geomechanics of coal-gas interactions: the role of coal ... · geomechanics. these include...
TRANSCRIPT
Geomechanics of Coal-Gas Interactions: The Role of Coal Permeability Evolution
by
Zhongwei Chen BEng
This thesis is presented for the degree of Doctor of Philosophy
of
The University of Western Australia
Petroleum Engineering School of Mechanical and Chemical Engineering
June 2012
School of Mechanical and Chemical Engineering
Abstract The University of Western Australia
I
Abstract
Complex interactions between stress and sorptive chemistry exert strong influence on coal
geomechanics. These include influences on gas sorption and flow, coal deformation,
porosity change and permeability modification. In this study, this chain of reactions is
labelled as ―coupled processes‖ implying that one physical process affects the initiation and
progress of another. The evolution of coal permeability is probably the most important
cross coupling to rigorously formulate the Geomechanics of coal-gas interactions.
There are an extensive suite of coal permeability models available in the literature – with
many of these models implemented into computer simulators to quantify coal-gas
interactions. The comparison of laboratory and field observations against the spectrum of
models indicate that current models have so far failed to explain the results from stress-
controlled shrinkage/swelling laboratorial tests and have only achieved some limited
success in explaining and matching in situ data. Almost all the permeability models are
derived for the coal as a porous medium, but used to explain the compound behaviours of
coal matrix and fracture. These review conclusions suggest that the impact of coal matrix-
fracture compartment interactions on the evolution of coal permeability has not yet been
understood well and further improvements are necessary. This knowledge gap defines the
goal of this study.
The issue as defined above has been addressed through an integrated approach of
experimental study, permeability model development, its implementation into a FE
simulator, and applications to field operations. Major findings are summarized as follows:
A series of gas flow-through experiments were conducted to quantify the sole influence
of the effective stress and sorption-induced strain on permeability evolution.
Experimental results demonstrate that the effective stress coefficient cannot be assumed
to be unity as generally done, and that the variations of the effective stress coefficient
School of Mechanical and Chemical Engineering
Abstract The University of Western Australia
II
under the influence of different gasses play an increasingly important role in the
determination of coal permeability evolution.
Based on the experimental observations, a phenomenological permeability model was
developed to quantify coal permeability reduction under the stress controlled conditions.
This new permeability model combines the effect of swelling strain with that of the
mechanical effective stress through the concepts of nature strain and partition ratio of
total swelling strain. A directional permeability model was also introduced to define the
evolution of gas sorption-induced permeability anisotropy under the full spectrum of
mechanical conditions. The model results are consistent with the experimental data or
field observations. Both models have been implemented into a fully coupled finite
element model to recover the important non-linear responses due to effective stress and
coal matrix shrinkage/swelling effects, where mechanical influences are rigorously
coupled with the gas transport system.
The role of heterogeneity on permeability evolution was then numerically conducted
under the unconstrained swelling conditions. The heterogeneous distributions of
Young's modulus and Langmuir strain constant in the vincity of the fracture were
included in the numerical model, and a net reduction of coal permeability was achieved
from the initial no-swelling state to the final equilibrium state and a good agreement was
obtained with laboratorial data under same conditions.
A dual-porosity/dual-permeability system was incorporated to investigate the effects of
coal properties, particularly the sorption time and shape factor, on the matrix-fracture
interactions for the dual-permeability system during CBM extraction. In this numerical
model, the complex multiphysics processes were coupled together with variable
permeability through the mass exchange term given as a function of sorption time,
pressure difference between matrix and fracture systems, matrix porosity, and matrix
particle volume. The sensitivity of the reservoir behavior to each parameter was
evaluated in details.
These results have been compiled into seven papers: four of them have been published in
International Journal of Coal Geology (IJCG) and International Journal of Greenhouse
Gas Control (IJGGC), one was submitted to 2011 Asia Pacific Coalbed Methane
Symposium, two are under review. This thesis is a compilation of these papers.
School of Mechanical and Chemical Engineering
Table of Contents The University of Western Australia
III
Table of Contents
ABSTRACT ....................................................................................................................................... I
TABLE OF CONTENTS .............................................................................................................. III
ACKNOWLEDGEMENTS .......................................................................................................... VII
PUBLICATIONS ARISING FROM THIS THESIS .................................................................. VIII
STATEMENT OF CANDIDATE CONTRIBUTION (%) ...........................................................XI
CHAPTER 1 .................................................................................................................................. 1-1
INTRODUCTION ................................................................................................................................................................ 1-1
1.1 BACKGROUND ....................................................................................................................................................... 1-1
1.2 PERMEABILITY MODEL DEVELOPMENTS ......................................................................................................... 1-3
1.2.1 PERMEABILITY MODELS UNDER UNIAXIAL STRAIN CONDITION ............................................. 1-3
1.2.2 PERMEABILITY MODELS UNDER VARIABLE STRESS CONDITIONS ........................................... 1-8
1.2.3 ANISOTROPIC PERMEABILITY MODELS ...................................................................................... 1-11
1.2.4 DUAL-POROSITY/DUAL-PERMEABILITY MODELS ..................................................................... 1-14
1.3 SIMULATION OF COUPLED MULTIPLE PROCESSES ...................................................................................... 1-14
1.3.1 ONE-WAY COUPLING ...................................................................................................................... 1-15
1.3.2 LOOSE COUPLING............................................................................................................................. 1-16
1.3.3 FULL COUPLING ................................................................................................................................ 1-18
1.4 RESEARCH GOALS .............................................................................................................................................. 1-19
1.5 THESIS OUTLINE .................................................................................................................................................. 1-20
1.6 REFERENCES......................................................................................................................................................... 1-21
CHAPTER 2 ................................................................................................................................. 2–1
LABORATORY CHARACTERISATION OF FLUID FLOW IN COAL WITH RESPECT TO GAS TYPE AND
TEMPERATURE ................................................................................................................................................................. 2–1
2.1 INTRODUCTION .................................................................................................................................................... 2–3
2.2 EXPERIMENTAL .................................................................................................................................................... 2–4
2.2.1 EXPERIMENT APPARATUS DESCRIPTION ..................................................................................... 2–4
2.2.2 ADSORPTION MEASUREMENT ....................................................................................................... 2–5
2.2.3 PERMEABILITY MEASUREMENT ..................................................................................................... 2–6
2.2.4 CLEAT COMPRESSIBILITY ................................................................................................................ 2–7
2.2.5 SWELLING MEASUREMENT ............................................................................................................. 2–8
School of Mechanical and Chemical Engineering
Table of Contents The University of Western Australia
IV
2.2.6 YOUNG’S MODULUS AND POISSON’S RATIO .............................................................................. 2–8
2.3 RESULTS AND DISCUSSION ............................................................................................................................... 2–9
2.3.1 ADSORPTION ISOTHERM RESULTS ................................................................................................ 2–9
2.3.2 SWELLING..........................................................................................................................................2–10
2.3.3 PERMEABILITY STRESS BEHAVIOUR ............................................................................................2–11
2.3.4 CLEAT COMPRESSIBILITY ...............................................................................................................2–12
2.3.5 GEOMECHANICAL PROPERTIES ....................................................................................................2–14
2.4 CONCLUSION ......................................................................................................................................................2–16
2.5 ACKNOWLEDGEMENT ......................................................................................................................................2–16
2.6 REFERENCES ........................................................................................................................................................2–17
CHAPTER 3 .................................................................................................................................. 3-1
LINKING GAS-SORPTION INDUCED CHANGES IN COAL PERMEABILITY TO DIRECTIONAL STRAINS
THROUGH A MODULUS REDUCTION RATIO ............................................................................................................ 3-1
3.1 INTRODUCTION ..................................................................................................................................................... 3-2
3.1.1 EXPERIMENTAL OBSERVATIONS .................................................................................................... 3-3
3.1.2 PERMEABILITY MODELS ................................................................................................................... 3-4
3.1.3 THIS STUDY ......................................................................................................................................... 3-5
3.2 APPROACH ............................................................................................................................................................. 3-6
3.2.1 COAL DEFORMATION ANALYSIS .................................................................................................... 3-7
3.2.2 FLOW AND TRANSPORT ANALYSIS ................................................................................................ 3-8
3.2.3 COAL PERMEABILITY ANALYSIS ..................................................................................................... 3-9
3.2.4 COUPLED MODEL ........................................................................................................................... 3-13
3.3 UNIAXIAL STRAIN CONDITION ..................................................................................................................... 3-14
3.4 DISPLACEMENT CONTROLLED CONDITION ............................................................................................... 3-16
3.5 FIELD CASE .......................................................................................................................................................... 3-17
3.6 EVALUATION OF COUPLED PROCESSES ....................................................................................................... 3-19
3.7 CONCLUSIONS .................................................................................................................................................... 3-21
3.8 ACKNOWLEDGEMENTS .................................................................................................................................... 3-22
3.9 REFERENCES ........................................................................................................................................................ 3-22
CHAPTER 4 .................................................................................................................................. 4-1
EFFECT OF THE EFFECTIVE STRESS COEFFICIENT AND SORPTION-INDUCED STRAIN ON THE EVOLUTION
OF COAL PERMEABILITY: EXPERIMENTAL OBSERVATIONS .................................................................................. 4-1
4.1 INTRODUCTION .............................................................................................................................................. 4-2
4.1.1 EXPERIMENTS ON COAL SWELLING/SHRINKAGE AND PERMEABILITY CHANGE ................... 4-2
4.1.2 MESUREMENTS OF EFFECTIVE STRESS COEFFICIENT ................................................................. 4-3
4.1.3 OBJECTIVE OF THIS STUDY ............................................................................................................... 4-6
School of Mechanical and Chemical Engineering
Table of Contents The University of Western Australia
V
4.2 EXPERIMENTAL METHODS ................................................................................................................................. 4-6
4.2.1 EXPERIMENT APPARATUS ................................................................................................................ 4-6
4.2.2 PERMEABILITY MEASUREMENT ...................................................................................................... 4-8
4.2.3 EXPERIMENTAL PROCEDURE .......................................................................................................... 4-8
4.2.4 WORK FLOW OF DATA ANALYSIS.................................................................................................... 4-9
4.3 RESULTS AND DISCUSSION ............................................................................................................................... 4-10
4.3.1 RESULTS FOR HELIUM INJECTION ................................................................................................ 4-10
4.3.2 RESULTS FOR METHANE INJECTION............................................................................................. 4-14
4.3.3 RESULTS FOR CARBON DIOXIDE INJECTION ............................................................................... 4-16
4.4 DISCUSSION ......................................................................................................................................................... 4-19
4.5 CONCLUSIONS ..................................................................................................................................................... 4-22
4.6 ACKNOWLEDGMENTS ....................................................................................................................................... 4-23
4.7 REFERENCES......................................................................................................................................................... 4-23
CHAPTER 5 .................................................................................................................................. 5-1
INFLUENCE OF THE EFFECTIVE STRESS COEFFICIENT AND SORPTION-INDUCED STRAIN ON THE
EVOLUTION OF COAL PERMEABILITY: MODEL DEVELOPMENT AND ANALYSIS .............................................. 5-1
5.1 INTRODUCTION ..................................................................................................................................................... 5-3
5.2 PERMEABILITY MODEL DEVELOPMENT ........................................................................................................... 5-5
5.2.1 EVALUATION OF EFFECTIVE STRESS EFFECTS .............................................................................. 5-6
5.2.2 EVALUATION OF SORPTION-INDUCED STRAIN EFFECTS ........................................................... 5-9
5.2.3 DEVELOPMENT OF COAL PERMEABILITY MODEL ..................................................................... 5-11
5.2.4 PHYSICAL MEANING OF SENSITIVITY RATIO .............................................................................. 5-12
5.3 PERMEABILITY MODEL EVALUATION ............................................................................................................ 5-13
5.3.1 PERMEABILITY MODEL VERIFICATION ........................................................................................ 5-14
5.3.2 COMPARISON WITH OTHER PERMEABILITY MODELS ............................................................... 5-16
5.4 MODEL IMPLEMENTATION .............................................................................................................................. 5-20
5.4.1 MODEL DESCRIPTIONS ................................................................................................................... 5-21
5.4.2 SIMULATION RESULTS AND ANALYSIS ........................................................................................ 5-23
5.5 CONCLUSIONS ..................................................................................................................................................... 5-27
5.6 ACKNOWLEDGMENTS ....................................................................................................................................... 5-29
5.7 REFERENCES......................................................................................................................................................... 5-29
CHAPTER 6 .................................................................................................................................. 6-1
ROLES OF COAL HETEROGENEITY ON EVOLUTION OF COAL PERMEABILITY UNDER UNCONSTRAINED
BOUNDARY CONDITIONS .............................................................................................................................................. 6-1
6.1 INTRODUCTION ..................................................................................................................................................... 6-2
6.2 THEORETICAL EVALUATION OF COAL PERMEABILITY MODELS ............................................................... 6-5
School of Mechanical and Chemical Engineering
Table of Contents The University of Western Australia
VI
6.2.1 GENERAL COAL PERMEABILITY MODEL ........................................................................................ 6-5
6.2.2 EVALUATION OF PERMEABILITY MODEL UNDER TWO BOUNDARY CONDITIONS ................ 6-6
6.3 A HETEROGENEOUS MATRIX-FRACTURE INTERACTION MODEL ........................................................... 6-11
6.3.1 NUMERICAL MODEL IMPLEMENTATION .................................................................................... 6-11
6.3.2 PERFORMANCE FOR A HOMOGENEOUS COAL .......................................................................... 6-13
6.3.3 PERFORMANCE FOR A HETEROGENEOUS COAL ....................................................................... 6-17
6.4 VERIFICATION WITH EXPERIMENTAL DATA ............................................................................................... 6-24
6.5 CONCLUSIONS .................................................................................................................................................... 6-26
6.6 ACKNOWLEDGEMENTS .................................................................................................................................... 6-26
6.7 REFERENCES ........................................................................................................................................................ 6-27
CHAPTER 7 .................................................................................................................................. 7-1
IMPACT OF VARIOUS PARAMETERS ON THE PRODUCTION OF COALBED METHANE ................................... 7-1
7.1 INTRODUCTION ..................................................................................................................................................... 7-3
7.2 METHODOLOGY .................................................................................................................................................... 7-6
7.2.1 COAL DEFORMATION ........................................................................................................................ 7-7
7.2.2 COAL PERMEABILITY MODEL ........................................................................................................... 7-9
7.2.3 MASS EXCHANGE FUNCTION .......................................................................................................... 7-9
7.2.4 SHAPE FACTOR FOR DUAL-PERMEABILITY MODEL ................................................................ 7-11
7.2.5 GAS FLOW EQUATIONS .................................................................................................................. 7-12
7.2.6 GAS PRODUCTION RATE ............................................................................................................... 7-13
7.3 EVALUATION OF GAS PRODUCTION ............................................................................................................. 7-15
7.3.1 MODEL DESCRIPTION .................................................................................................................... 7-15
7.3.2 MODELLING STRATEGY ................................................................................................................. 7-16
7.3.3 RESERVOIR BEHAVIOUR UNDER CONSTANT TOTAL PRODUCTIVE GAS VOLUME ............. 7-17
7.3.4 RESERVOIR BEHAVIOUR UNDER VARIABLE TOTAL PRODUCTIVE GAS VOLUME ............... 7-21
7.4 FIELD APPLICATION OF DUAL-PERMEABILITY MODEL ............................................................................. 7-27
7.5 CONCLUSION ...................................................................................................................................................... 7-29
7.6 ACKNOWLEDGEMENTS .................................................................................................................................... 7-30
7.7 REFERENCES ........................................................................................................................................................ 7-31
CHAPTER 8 .................................................................................................................................. 8-1
CONCLUDING REMARKS ............................................................................................................................................... 8-1
8.1 MAIN FINDINGS .................................................................................................................................................... 8-1
8.2 RECOMMENDATIONS FOR FUTURE WORK ...................................................................................................... 8-4
School of Mechanical and Chemical Engineering
Acknowledgements The University of Western Australia
VII
Acknowledgements
I would like to express my deep and sincere gratitude to my supervisor, Professor Jishan
Liu, who invited me to Australia and supported me persistently during the period of this
research. Prof. Liu was always there to listen and give advice, which enabled my research
work to move forward continuously. Many of the ideas in this thesis would not have taken
shape without his incisive thinking and insightful suggestions. What I learned from him will
benefit me greatly in the rest of my life.
Many thanks go to my Co-supervisors Dr. Zhejun Pan and Dr. Luke D. Connell from
CSIRO for providing experimental facilities, financial support, and critical and insightful
suggestions of some of the papers involved in this thesis. Special thanks to Professor
Derek Elsworth from Pennsylvania State University, who has given me continued and
invaluable support and help in my study and life.
I am indebted to the research group members from School of Mechanical and Chemical
Engineering for their friendship and diverse help during my study, including Dr. Jianguo
Wang, Dr. Jian Li, Dr. Jianxin Liu, Dr. Yu Wu, Mr. Hamid Ghafram Al Shahri, Mr. Dong
Chen and Ms. Hongyan Qu. I have also loved working with colleagues from CSIRO Earth
Science and Resource Engineering (Clayton), and thanks to Dr. Meng Lu, Mr. David
Down, Mr. Greg Lupton, Mr. Nick Lupton and Mr. Michael Camilleri.
I would like to acknowledge the Scholarship for International Research Fees of University
of WA (SIRF), Western Australia- CSIRO postgraduate Scholarship (WACUPS), and
CSIRO National Flagship Top-Up Scholarship for providing the financial support to me to
pursue this study.
At last, I wish to express my sincere gratitude to my father Chunfu Chen, my mother Suqin
Zou, and my wife Nana Wang, for their constant love and inspiration. Without their
support, I could not have done it.
School of Mechanical and Chemical Engineering
Publications Arising from This Thesis The University of Western Australia
VIII
Publications Arising From This Thesis
Journal papers
1. Chen, Z., Liu, J., Pan, Z., Connell, L.D., Elsworth, D., 2012 Effect of the Effective
Stress Coefficient and Sorption-Induced Strain on the Evolution of Coal
Permeability: Model development and analysis. International Journal of
Greenhouse Gas Control, 8, 101-110.
2. Zheng, G, Pan, Z., Chen, Z., Tang S., 2012. Laboratory study of gas permeability
and cleat compressibility for CBM/ECBM in Chinese Coals. Energy Exploration &
Exploitation 30(3), 451-476.
3. Qu, H., Liu, J., Chen, Z., Wang, J., 2012. Complex evolution of coal permeability
during CO2 injection under variable temperatures. International Journal of
Greenhouse Gas Control 9, 281-293.
4. Wang, J., Liu, J., Kabir, A., Chen, Z., 2012. Effects of Non-Darcy Flow on the
Performance of Coal Seam Gas Wells. International Journal of Coal Geology93, 62-
74.
5. Chen, Z., Pan, Z., Liu, J., Connell, L.D., Elsworth, D., 2011. Effect of the effective
stress coefficient and sorption-induced strain on the evolution of coal permeability:
Experimental observations. International Journal of Greenhouse Gas Control 5,
1284-1293.
6. Liu, J., Chen, Z., Elsworth, D., Qu, H., Chen, D., 2011. Interactions of multiple
processes during CBM extraction: A critical review. International Journal of Coal
Geology 87, 175-189.
7. Liu, J., Chen, Z., Elsworth, D., Miao, X., Mao, X., 2011. Evolution of coal
permeability from stress-controlled to displacement-controlled swelling conditions.
Fuel 90, 2987-2997.
8. Liu, J., Wang, J., Chen, Z., Wang, S., Elsworth, D., Jiang, Y., 2011. Impact of
transition from local swelling to macro swelling on the evolution of coal
permeability. International Journal of Coal Geology 88, 31-40.
9. Wu, Y., Liu, J., Chen, Z., Elsworth, D., Pone, D., 2011. A dual poroelastic model
for CO2-enhanced coalbed methane recovery. International Journal of Coal
Geology 86, 177-189.
School of Mechanical and Chemical Engineering
Publications Arising from This Thesis The University of Western Australia
IX
10. Liu, J., Chen, Z., Elsworth, D., Miao, X., Mao, X., 2010. Linking gas-sorption
induced changes in coal permeability to directional strains through a modulus
reduction ratio. International Journal of Coal Geology 83, 21-30.
11. Liu, J., Chen, Z., Elsworth, D., Miao, X., Mao, X., 2010. Evaluation of stress-
controlled coal swelling processes. International Journal of Coal Geology 83, 446-
455.
12. Chen, Z., Liu, J., Elsworth, D., Connell, L.D., Pan, Z., 2010. Impact of CO2
injection and differential deformation on CO2 injectivity under in-situ stress
conditions. International Journal of Coal Geology 81, 97-108.
13. Wu, Y., Liu, J., Elsworth, D., Chen, Z., Connell, L., Pan, Z., 2010. Dual poroelastic
response of a coal seam to CO2 injection. International Journal of Greenhouse Gas
Control 4, 668-678.
Selected conferences papers
1. Pan, Z., Chen, Z., Connell, L.D., Lupton, N.. Laboratory Characterization of Fluid
Flow in Coal for Different Gasses at Different Temperatures. 2011 Asia Pacific
Coalbed Methane Symposium, Brisbane, Australia.
2. Zhou, L., Feng, Q., Chen, Z., Liu, J., 2011. Modeling and Upscaling of Binary Gas
Coal Interactions in CO2 Enhanced Coalbed Methane Recovery. International
Conference on Environment Science and Engineering (ICESE 2011), Bali Island,
Indonesia.
3. Liu, J., Chen, Z., Qu, H., 2010. Multiphysics of Coal-Gas Interactions: The
Scientific Foundation for CBM Extraction. Asia Pacific Oil & Gas Conference and
Exhibition, Brisbane, Australia. SPE-133015-PP.
4. J.G. Wang, Liu, J., Liu, J., Chen, Z., 2010. Impact of rock microstructures on the
supercritical CO2 enhanced gas Recovery. International Oil & Gas Conference and
Exhibition, Beijing China. SPE-131759-PP.
5. Chen, Z., Liu, J., Pan, Z., Connell, L.D., Elsworth, D., 2010. Relations between
Coal Permeability and Directional Strains and Their Application to San Juan Basin.
44th U.S. Rock Mechanics Symposium and 5th U.S.-Canada Rock Mechanics
Symposium, Salt Lake City, Utah, USA. ARMA 10-245.
6. Chen, Z., Liu, J., Elsworth, D., Connell, L.D., Pan, Z., 2009. Investigation of CO2
Injection Induced Coal-Gas Interactions. The 44th U.S. Rock Mechanics
School of Mechanical and Chemical Engineering
Publications Arising from This Thesis The University of Western Australia
X
Symposium and 5th U.S.-Canada Symposium, Asheville, North Carolina, USA.
ARMA-09-099.
7. Wu, Y., Liu, J., Elsworth, D., Chen, Z., Connell, L., Pan, Z., 2009. Dual
poroelastic responses of coal to CO2 injection. The 44th U.S. Rock Mechanics
Symposium and 5th U.S.-Canada Symposium, Asheville, North Carolina, USA. No.
ARMA-09-164.
8. Chen, Z., Liu, J., Elsworth, D., Pan, Z., Connell, L.D., 2009. In-situ numerical
testing of CO2 sequestration in coal: Effects of confining stress and injection
pressure. 2009 Asia Pacific Coalbed Methane Symposium, Xuzhou, China.
9. Wu Y., Liu J., Chen Z., Elsworth D., Connell, L.D., 2009. In-situ numerical testing
of CO2 sequestration in coal: Dual poroelastic effects. 2009 Asia Pacific Coalbed
Methane Symposium, Xuzhou, China.
10. Liu J., Chen Z., Wu Y., Elsworth D., 2009. Multiphysics of Coal-Gas Interactions.
Seventh International Symposium on Rockburst and Mine Seismicity, Dalian,
China.
11. Chen, Z., Liu, J., Connell, L., Pan, Z., 2008. Impact of Effective Stress and CH4-
CO2 Counter-Diffusion on CO2 Enhanced Coalbed Methane Recovery. 2008 Asia
Pacific Oil & Gas Conference and Exhibition, Perth, Australia. SPE 116515.
School of Mechanical and Chemical Engineering
Statement of Candidate Contribution (%) The University of Western Australia
XI
Statement of Candidate Contribution (%)
This thesis contains published work and/or work prepared for publication, which has
been co-authored. The bibliographical details of the work and where it appears in the
thesis are outlined below.
Liu, J. (30%), Chen, Z. (45%), Elsworth, D. (15%), Qu, H. (5%) and Chen, D. (5%), 2011.
Interactions of multiple processes during CBM extraction: A critical review. International
Journal of Coal Geology, 87(3-4): 175-189.
Pan, Z. (30%), Chen, Z. (40%), Connell, L.D. (20%) and Lupton, N. (10%). Laboratory
Characterisation of Fluid Flow in Coal for Different Gasses at Different Temperatures,
Asia Pacific Coalbed Methane Symposium, Brisbane, Australia, in May, 2011.
Liu, J. (30%), Chen, Z. (50%), Elsworth, D. (10%), Miao, X. (5%) and Mao, X. (5%), 2010.
Linking gas-sorption induced changes in coal permeability to directional strains through a
modulus reduction ratio. International Journal of Coal Geology, 83(1): 21-30.
Chen, Z. (70%), Pan, Z. (10%), Liu, J. (10%), Connell, L.D. (5%) and Elsworth, D. (5%),
2011. Effect of the effective stress coefficient and sorption-induced strain on the evolution
of coal permeability: Experimental observations. International Journal of Greenhouse Gas
Control, 5(5): 1284-1293.
Chen, Z. (70%), Liu, J. (10%), Pan, Z. (10%), Connell, L.D. (5%) and Elsworth, D. (5%).
Effect of the effective stress coefficient and sorption-induced strain on the evolution of
coal permeability: Model development and analysis. International Journal of Greenhouse
Gas Control (Accepted).
Chen, Z. (60%), Liu, J. (20%), Elsworth, D. (5%), Wang, J. (5%), Pan, Z. (5%), and
Connell, L.D. (5%). Roles of Coal Heterogeneity on Evolution of Coal Permeability under
Unconstrained Boundary Conditions.
Chen, Z. (50%), Kabir, A. (20%), Liu, J. (15%), Wang, J. (5%), Pan, Z. (5%), and Connell,
L.D. (5%). Impact of Various Parameters on the Production of Coalbed Methane.
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-1
Chapter 1
Introduction
Jishan Liua, Zhongwei Chena, Derek Elsworthb, Hongyan Qua, Dong Chena
a School of Mechanical and Chemical Engineering, the University of Western Australia, WA 6009, Australia
b Department of Energy and Mineral Engineering, Penn State University, PA 16802-5000, USA
1.1 Background
Advances in our understanding of coal-gas interactions have changed the manner in which
we treat coalbed methane: from mitigating its dangers as a mining hazard to developing its
potential as an unconventional gas resource recovered as a useful by-product of CO2
sequestration.
As found in nature, coal is a typical dual porosity/permeability system (Harpalani and
Schraufnagel, 1990; Lu and Connell, 2007; Warren and Root, 1963) containing porous
matrix surrounded by fractures. These natural fractures form a closely-spaced, orthogonal
network called cleats. The main set of fractures, termed face cleats, is comprised of well-
developed, extensive, roughly planar fractures that run parallel to one another. Butt cleats
are orthogonal to face cleats and often terminate at them. Butt cleats are also roughly
planar but are not as well-developed or as continuous as face cleats. The cleat system
provides an essential and effective flow path for gas. Much of the measured bulk or ―seam‖
permeability is due to the cleat system, although the presence of larger scale discontinuities
such as fractures, joints, and faults can also make a significant contribution. The coal
matrix is isolated by the fracture network and is the principal medium for storage of the gas
(of the order of 98%), principally in adsorbed form and with low permeability in
comparison to the bounding cleats (Gray, 1987). The remaining gas is stored in the natural
fractures, or cleats, either as free gas or dissolved in water. The surface area of the coal on
which the methane is adsorbed is very large (20 to 200 m2/g) (Patching, 1970) and gas is
stored at near-liquid densities.
Coalbed methane has shown the enormous potential in the past several decades. Major
reserves exist in many countries and more than 90% of the estimated reserves are in
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-2
Canada, Russia, China, the United States and Australia. Kuuskraa et al. (1992) defined
global CBM reserves through a detailed study of coalbed basins around the world. This
work was further updated by different researchers before the final form was presented
(Boyer, 1994; Kuuskraa et al., 1992; Murray, 1996; Palmer, 2008; White et al., 2005).
Estimates of the global coalbed methane (CBM) reserve defined in volume of CH4 are
summarized in Table 1-1.
Table 1-1. Coalbed methane reserves around the world
Country CBM Reserves (Tcf)
Boyer (1994) Murray (1996) Kuuskraa (1998) Palmer (2008)
Canada 200-2,700 300-4,260 570-2,280 200-2,700
Russia 600-4,000 600-4,000 550-1,550 600-4,000
China 1,060-1,240 1,060-2,800 350-1,150 1,060-1,240
United States 343-414 275-650 500-1,730 400
Australia 300-500 300-500 310-410 300-500
Indonesia - - 210 -
Germany 100 100 120 (Western Europe) 100
Poland 100 100 70 100
United Kingdom 60 60 - 60
Ukraine 60 60 50 60
Kazakhstan 40 40 40 40
Southern Africa* 30 40 100 30
India 30 30 90 30
Turkey - - 50 -
Total 2,953-9,304 2,976-12,640 3,010-7,840 2,980-9,260
*Includes South Africa, Zimbabwe and Botswana.
Compared to conventional gas reservoirs, coal reservoirs have low effective porosity and
high compressibility and are dominated by gas desorption. CBM recovery triggers a series
of coal-gas interactions. For primary gas production, the reduction of gas pressure
increases effective stress which in turn closes fracture apertures and reduces the
permeability. As the gas pressure reduces below the desorption point, methane is released
from the coal matrix to the fracture network and the coal matrix shrinks. As a direct
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-3
consequence of this matrix shrinkage the fractures may dilate (zero volume change
condition) and fracture permeability correspondingly increases. Thus a rapid initial
reduction in fracture permeability (due to an increase in effective stress) is supplanted by a
slow increase in permeability (indexed to matrix shrinkage). Whether the ultimate, long-
term, permeability is greater or less than the initial permeability depends on the net
influence of these dual competing mechanisms (Chen et al., 2008; Connell, 2009; Liu et al.,
2010b,c; Shi and Durucan, 2004). Therefore, understanding the transient characteristics of
permeability evolution in fractured coals is of fundamental importance to CBM recovery.
CBM extraction induced complex interactions between stress and sorptive chemistry exert
strong influence on the transport and sorptive properties of the coal. These include
influences on gas sorption and flow, coal deformation, porosity change and permeability
modification. We label this chain of reactions as ―coupled processes‖ implying that one
physical process affects the initiation and progress of another. The individual processes, in
the absence of full consideration of cross couplings, form the basis of very well-known
disciplines such as elasticity, hydrology and heat transfer. Therefore, the inclusion of cross
couplings is the key to rigorously formulate the behaviour for coupled processes of coal-
gas interactions. The complexity of these interactions is reflected in the extensive suite of
coal permeability models available in the literature– with many of these models
implemented into computer simulators to quantify coal-gas interactions. The primary goal
of this paper is to review and evaluate the performance of these disparate models of coal
permeability evolution and to define principal physical conditions where they, and their
application in simulators, can be most successful.
1.2 Permeability model developments
1.2.1 Permeability models under uniaxial strain condition
An equation for permeability and porosity of a collection of matchsticks are discussed by
Reiss (1980) together with an equation for collections of slabs and cubes. As the coal
deposit is idealized as a collection of matchsticks, flow in the core sample is along the axis
of the matchsticks. Permeability for this geometry is given by Reiss (1980) as
32
48
1fak (1-1)
where a is cleat spacing, and f is cleat porosity.
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-4
Differentiating with respect to hydrostatic stress and combining the relationship between
coal physical properties gives (Seidle et al., 1992; Pan et al., 2010)
f
h
cE
kk
3212
(1-2)
where fc is coal cleat compressibility defined as
p
f
f
1,
h is the horizontal stress. E
and are coal Young's Modulus and Poisson's ratio respectively.
The first term in parentheses represents the contribution of changes in the coal matrix to
the stress dependence of permeability. This term is analogous to grain compaction in
conventional reservoir rocks. The second term represents the cleat volume contribution to
the stress-permeability relation, which is normally two to three orders of magnitude larger
than the coal matrix term (Reiss, 1980). Therefore, simplifying and integrating the above
equation gives
0
0
3exp hhfck
k (1-3)
where 0h is the initial horizontal stress.
Gray (1987) considered the changes in the cleat permeability to be primarily controlled by
the prevailing effective horizontal stresses that act across the cleats. Under the assumption
of uniaxial strain, the influence of matrix shrinkage on changes in coal permeability was
first incorporated into a permeability model. The horizontal stress incorporating matrix
shrinkage was expressed as
s
s
shh p
p
Epp
1100 (1-4)
where sp refers to equivalent sorption pressure.
By assuming that an individual fracture reacts as an elastic body upon a change in the
normal stress component, Gilman and Beckie (2000) proposed a simplified mathematical
model of methane movement in a coal seam taking into account the following features: a
relatively regular cleat system, adsorptive methane storage, an extremely slow mechanism
of methane release from the coal matrix into cleats and a significant change of permeability
due to desorption. Using the uniaxial strain assumption and Terzaghi formula, the effective
stress in horizontal plane, h
e , was expressed as below, which is similar to Gray's (1987)
result:
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-5
SE
ph
e
11 (1-5)
where S is the change of the adsorbate mass and is the volumetric welling/shrinkage
coefficient.
The exponential relation was used for the permeability calculation
S
Ep
EEk
k
ff
h
e
11
3exp
3exp
0
(1-6)
where fE is an analogous Young's modulus for the fracture.
Seidle and Huitt (1995) calculated the permeability increase due to matrix shrinkage alone
by assuming that coal sorption-induced strain is proportional to the amount of gas sorbed
and that the sorbed gas is related to pressure by Langmuir's equation. Their porosity and
permeability models were defined as
pP
p
pP
p
LL
L
0
0
00
21
31
(1-7)
3
0
0
00
21
31
pp
p
pp
p
k
k
LL
L
(1-8)
where L and Lp are the maximum volumetric strain and gas pressure at which the matrix
strain is half of the maximum value, respectively.
This model considered the effects of coal-matrix swelling/shrinkage only, ignoring the
impact of coal compressibility. Therefore, their model is limited to specific conditions in
which sorption-induced strain (matrix swelling or shrinkage) dwarfs pressure-induced,
elastic changes in cleat permeability (Robertson 2005).
Based on the matchstick geometry model and the relation between permeability and
porosity developed by Seidle and Huitt (1995), Shi and Durucan (2004) presented a model
for pore pressure-dependent cleat permeability for gas-desorbing, linear elastic coalbeds
under uniaxial strain conditions. In this model, it was assumed that changes in the cleat
permeability of coalbeds were controlled by the prevailing effective horizontal stresses
normal to the cleats. Variations in the effective horizontal stresses under uniaxial strain
conditions are expressed as a function of pore pressure reduction during drawdown, which
includes a cleat compression term and a matrix shrinkage term that have competing effects
on cleat permeability, as expressed below
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-6
LL
Lhhpp
p
pp
pEpp
0
000
131
(1-9)
00 3exp hhfckk (1-10)
Based on the theory of linear elasticity for strain changes, Palmer and Mansoori (1996)
developed another widely used theoretical coal permeability model as a function of pore
pressure and matrix shrinkage. In this model, the incremental pore volume strain, pd , can
be defined as a result of a simple volumetric balance between the bulk rock, the grains, and
the pores
gr
p dd
d
1 (1-11)
where rd is the incremental rock volume strain, gd and are incremental grain volume
strain and porosity, respectively.
By assuming the uniaxial strain condition, 1 , and no change in overburden stress
results in
0
0
0
0
00
11pP
p
pP
p
M
Kpp
c
LL
Lm
(1-12)
The cubic relation between porosity and permeability was used for this derivation, as
shown below
3
0
0
0
0
00
11
pP
p
pP
p
M
Kpp
c
k
k
LL
Lm
(1-13)
where '1
1
f
M
K
Mcm ,
211
1
M , ' is grain compressibility and f is a
fraction between 0 and 1.
An improved P-M model has been developed, and is summarized in Palmer et al. (2007).
The model now includes (1) cleat anisotropy and potential suppression of pressure-
dependent permeability, (2) modulus changes with depletion, and (3) undersaturated coals.
Similarly, the Advanced Resources International (ARI) group developed another
permeability model (Pekot and Reeves, 2002). This model does not have a geomechanics
framework, but instead extracts matrix strain changes from a Langmuir curve of strain
versus reservoir pressure, which is assumed to be proportional to the gas concentration
curve. The matrix shrinkage is proportional to the adsorbed gas concentration change,
multiplied by shrinkage compressibility mC (a free parameter). The ARI model has been
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-7
compared to the P-M model, and the conclusion was that the two models are essentially
equivalent in saturated coals, and where the strain versus pressure function is proportional
to the Langmuir isotherm (Palmer et al., 2007).
Following the above work, Cui and Bustin (2005) investigated quantitatively the effects of
reservoir pressure and sorption-induced volumetric strain on coal-seam permeability with
constraints from the adsorption isotherm and associated volumetric strain and derived a
stress-dependent permeability model. Initially the authors used poroelasticity to achieve the
relation between porosity change and effective stress change, as shown below
00
0
1exp pp
K p
(1-14)
where pK is the bulk modulus for pore system.
The cubic relation between permeability and porosity was used to calculate coal cleat
permeability change.
00
3
00
3exp pp
Kk
k
p
(1-15)
By assuming constant overburden stress and uniaxial strain conditions, then this
permeability model was extended to be
00
0 19
2
13
13exp ss
p
Epp
Kk
k
(1-16)
Pan and Connell (2007) developed a theoretical model for sorption-induced strain and
applied to single-component adsorption/strain experimental data. Clarkson (2008)
expanded this theoretical model to calculate the sorption-strain component of the P-M
model (Palmer et al., 2007). The expressions for sorption-induced strain and permeability
calculation are given as
s
s
s
s
ss
E
pxf
EpBRTL
21,)1ln( (1-17)
3
0
0
00
11
1
s
m
M
Kpp
c
k
k
(1-18)
wheresE is the modulus of the solid phase,
sv is Poisson's ratio for solid phase and s is
the density for solid phase. R is the gas constant (8.314J mol-1K-1), T is the temperature
(K), L is Langmuir sorption constant (mol/kg), and B is Langmuir pressure constant (Pa-1).
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-8
1.2.2 Permeability models under variable stress conditions
Robertson and Christiansen (2006) described the derivation of a new equation that can be
used to model the permeability behaviour of a fractured, sorptive-elastic medium, such as
coal, under variable stress conditions. The model is derived for cubic geometry rather than
matchstick geometry under biaxial or hydrostatic confining pressures, and it is also
designed to handle changes in permeability caused by adsorption and desorption of gasses
from the matrix blocks.
In this model, the effective porosity of the matrix block is assumed to be zero, leaving the
fracture system to provide the only interconnected void space. The permeability model was
expressed as
00
0
00
In3
)(219)0(exp1
3exppP
pP
pP
Ppp
E
ppc
k
k
L
L
L
LL
c
cf
(1-19)
where c is the change rate in fracture compressibility.
Based on the theory of poroelasticity, a general porosity and permeability model was
developed by Zhang et al., (2008), where the expression of permeability for the pore system
is defined as
3
0
0
0
0
11
1
SSS
Sk
k
m
m
(1-20)
where s
s
vK
pS and 0
00 s
s
vK
pS .
Similarly, Connell et al. (2010) presented two new analytical permeability model
representations for standard triaxial strain and stress conditions, derived from the general
linear poroelastic constitutive law, that include the effects of triaxial strain and stress for
coal undergoing gas adsorption induced swelling. A novel approach is presented to
distinguish between the sorption strain of the coal matrix, the pores (or cleats) and the bulk
coal.
Contrary to previous models developed for field conditions, their model mainly deals with
variable stress conditions commonly used during measurement of permeability in the
laboratory.
When experimental results from these tests are interpreted, a matchstick or cubic coal
model is typically assumed with the matrix blocks completely separated from each other in
a stacked structure. Under this assumption, matrix swelling will not affect coal fracture
permeability under conditions of constant confining (total) stress, because, for a given pore
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-9
pressure, p , the coal matrix swelling will result in block swelling, rather than changes in
fracture aperture (Liu and Rutqvist, 2010, Liu, et al., 2011; Connell et al., 2010). The
effective stress is also decoupled from matrix swelling due to the complete separation
between matrix blocks caused by through-going fractures. Therefore, the permeability
should not change, but this is not consistent with laboratory observations (Harpalani and
Chen, 1997; Pini et al., 2009; Pan et al., 2010), which show dramatic reduction in
permeability with the injection of an adsorbing gas. Liu and Rutqvist (2010) believed that in
reality coal matrix blocks are not completely separated from each other by fractures but
connected by the coal-matrix bridges, and developed a new coal-permeability model, which
explicitly considers fracture–matrix interaction during coal-deformation processes based on
the internal swelling stress concept. For example, the effective stress under uniaxial strain
conditions can be calculated by the following equations
inse
EP
11 (1-21)
efC
fin eE
112
0 (1-22)
where in is the internal swelling stress, and
0f is the fracture porosity. The above
coupled equations are solved to obtain the effective stress and strain.
An alternate reasoning has been applied by Liu et al. (2010a) on this issue, considering that
the reason for the above phenomena may be the internal actions between coal fractures
and matrix have not been taken into consideration. A model capable of replicating this
apparently anomalous behaviour is developed by considering the interactions of the
fractured coal mass where cleats do not create a full separation between adjacent matrix
blocks, but where solid rock bridges are present. The role of swelling strains is
accommodated both over contact bridges that hold cleat faces apart but also over the non-
contacting span between these bridges. The effects of swelling act competitively over these
two components: increasing porosity and permeability due to swelling of the bridging
contacts but reducing porosity and permeability due to the swelling of the intervening free-
faces.
The fracture permeability was expressed as
3
0
3
0
)1(11
sv
f
m
f
f R
b
b
k
k
(1-23)
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-10
where b and b are fracture aperture and fracture aperture change, respectively. v is the
volumetric strain.
This study also considered the resultant change in coal permeability, which combined the
outcome of the reduction in fracture opening due to coal matrix swelling and effective
stress change and the decrease in effective stress due to changes in fluid pressure and
confining stress for the matrix system, as
3
000
0
3
000
0
0
111
sv
f
m
fm
fm
m
m
fm
m R
kk
k
K
pR
kk
k
k
k
(1-24)
where 0f is initial fracture porosity, and
mR is elastic modulus reduction ratio, defined as
mEE . mE is Young's modulus for coal matrix.
0mk and 0fk are initial coal matrix
permeability and coal fracture permeability respectively. Subscripts m and f refer to matrix
and fracture system respectively.
Izadi et al. (2011) proposed a mechanistic representation of coal as a collection of
unconnected cracks in an elastic swelling medium. The cracks are isolated from each other
but swelling within a homogeneous but cracked continuum results in a reduction in crack
aperture with swelling, and a concomitant reduction in permeability. In the limit, this
behaviour reduces to a change in permeability defined as a fully constrained model (zero
volume change) as,
3
0
23
0
3
0
11
L
Ls
pp
p
lb
s
lb
s
k
k (1-25)
where l is the crack length, s the cleat spacing and b0 the initial aperture and eL is the
Langmuir strain coefficient.
Ma et al. (2011) developed a model, which was based on the volumetric balance between
the bulk coal, and solid grains and pores, using the constant volume theory (Massarotto et
al., 2009). It incorporates primarily the changes in grain and cleat volumes and is, therefore,
different from the other models that lay heavy emphasis on the pore volume/cleat
compressibility. In this study, the overall matchstick strain resulting from matrix shrinkage
and decrease in pressure is given as
0
0
0 111 pp
Epp
p
pp
p
a
a
LL
L
(1-26)
The permeability change can be calculated by the following expression
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-11
a
a
a
a
k
k
1
21
3
0
0
(1-27)
where a and a are the matrix width and width change, respectively. 0 is the porosity at
virgin reservoir pressure.
1.2.3 Anisotropic Permeability Models
The permeability models as reviewed above do not reflect the directional behaviour of
permeability change. The anisotropic characteristics of a coal matrix-fracture structure
suggest that the evolution of coal permeability should be direction-dependent. With cubic
coal cores, Pomeroy and Robinson (1967) found that the flow rates of water
(corresponding to permeability) were significantly different when the confining pressures
were perpendicular to main cleats (face cleats), cross cleats (butt cleats) or bedding planes.
From field well tests, Koenig and Stubbs (1986) reported the anisotropy ratio of
permeability in the plane of bedding was as high as 17:1 in the Rock Creek coalbeds of the
Warrior Basin of the USA. Permeability anisotropy of coal was also confirmed by other
experimental results of Gash et al (1992). Using coal samples from the San Juan Basin and
under a confining stress of 6.9 MPa (1000 psi) they found that the permeability parallel to
bedding planes was 0.6~1.7 mD in the direction of the face cleat and 0.3~1.0 mD in the
direction of the butt cleat, but only 0.007 mD in the direction vertical to the bedding planes.
A few of coal permeability models have been developed to accommodate the anisotropy, as
summarized below.
Wong (2003) developed a model for deformable granular media, which quantifies the
anisotropic changes in permeability when the material experiences shear deformation. In
this study, the directions of the principal permeability magnitudes are governed by the
induced strains, so the effects of stress paths and stress levels are implicitly considered
through effective stress-strain constitutive laws. This strain-induced permeability model is
written as:
3
2
1
'''
'''
'''
30
20
10
3
2
1
abb
bab
bba
k
k
k
k
k
k
(1-28)
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-12
where, 'a and 'b are material constants and can be experimentally measured; 0ik and
ik
denote initial and current permeability, respectively; i are current principal strains.
Following the above work, Al-Yousef (2005) presented an analytical solution for the
steady-state flow problem for anisotropic permeability measurements. Gu and Chalaturnyk
(2005) developed another permeability model. In this model, coalbeds are considered as
naturally fractured reservoirs, and represented with a collection of matchsticks. The
permeability is expressed as
3
0
1
li
i
b
a
k
k (1-29)
lTilDilPilEili (1-30)
where li is the directional effective strain and each term represents in order the
mechanical deformation due to stress change, the mechanical deformation due to pressure
change, matrix shrinkage/swelling due to desorption/sorption, and thermal
contract/expansion due to temperature changes.
Recently, they extended their work by considering discontinuous coal masses as an
equivalent elastic continuum. The implementation procedure of an explicit-sequential
coupled simulation using such permeability models in industrial simulators is complex but
feasible for coupled simulation in pressure depleting CBM reservoirs (Gu and Chalaturnyk,
2010). The total change of cleat aperture is defined as
3.0/uu tan
3.0/uu
ss
ss
pm
ff
p
f
ma
ab
(1-31)
The total change of matrix block is defined as
3.0/uu tan
3.0/uu
ss
ss
pt
Lm
m
ff
t
L
pt
Lmf
t
L
ba
baa
(1-32)
The following expression is used for the permeability calculation
j
j
n
jm
jm
i
i
a
a
b
b
k
k
j
1
1
3
0
(1-33)
where a is the width of the coal matrix block, f is the change of normal strain within the
fracture (cleat), su and p
su are shear displacements of the fracture and peak shear
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-13
displacement of the fracture respectively. f is change of the shear strain of a fracture,
and m is the mobilized dilation angle. t
L is the total change of linear strain of a
composite unit including a matrix block and a fracture, and mb is the mechanical aperture
of a fracture.
Wang et al. (2009) developed a model that incorporates the anisotropic structural and
mechanical properties to describe the directional permeability of coal. In this model, the
mechanical and non-mechanical deformations of coal under confined stress conditions that
imitate coal reservoirs are taken into account. The mechanical deformation is the stress-
dominated deformation that can be described using the general stress–strain correlation
and nonmechanical deformation is sorption-induced matrix swelling/shrinkage that was
treated using a thermal expansion/contraction analogy. A strain factor, depended on coal
properties and sorption characters such as coal type and rank, and sorbent gas, was
introduced to correct the strains theoretically obtained for better interpretations of
laboratory strain data under unconstrained conditions that are widely used for tests of coal
permeability.
Liu et al. (2010b) developed a permeability model to define the evolution of gas sorption-
induced permeability anisotropy under the full spectrum of mechanical conditions
spanning prescribed in-situ stresses through constrained displacement. In the model, gas
sorption-induced coal directional permeabilities are linked into directional strains through
an elastic modulus reduction ratio, which represents the partitioning of total strain for an
equivalent porous coal medium between the fracture system and the matrix. Verification of
this model has been conducted by et al. (2010b).
The directional permeability expression is defined as follows
ji
ej
f
m
i
i R
k
k3
00
)1(31
2
1
(1-34)
where 0f is the initial fracture porosity at reference conditions, zyxji ,,, .
Recently, Pan and Connell (2011) developed an anisotropic swelling model based on the
Pan and Connell (2007) swelling model, which applies an energy balance approach where
the surface energy change caused by adsorption is equal to the elastic energy change of the
coal solid. This new model also incorporated anisotropic coal properties.
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-14
1.2.4 Dual-porosity/dual-permeability models
Dual-permeability or multiple permeability models have been developed to represent the
porosity and permeability of all constituent components (Bai et al., 1993), including the role
of sorption (Bai et al., 1997), and of multiple fluids (Douglas et al., 1991). Moreover,
several models have been applied to represent the response of permeability evolution in
deforming aquifers and reservoirs (Elsworth and Bai, 1992; Ouyang and Elsworth, 1993;
Bai et al., 1995; Liu and Elsworth, 1997), to accommodate gas flow and other mechanical
influences (Zhao et al., 2004).
Wu et al. (2010a) developed a dual poroelastic model (dual solid media – coal matrix and
fracture) for single gas under variable stress conditions. The model allows exploration of
the full range of mechanical boundary conditions from invariant stress to restrained
displacement. Wu et al. (2010b) extended their previous work (Wu et al., 2010a) to define
the evolution of gas sorption-induced anisotropic permeability. In this study, dual
permeabilities are used which is different from Gu and Chalaturnyk's work (2010). The
expression of anisotropic permeability for cleat system is defined as
ji
eisTf
fi
i
KT
K
Kk
k
3
00
1
3
1
3
1
3
11
2
1
(1-35)
where 0f is the initial fracture porosity at reference conditions, zyxji ,,, . T ,
s ,
ei refer to the change in temperature, sorption-induced strain and mechanical effective
stress. The permeability model for the matrix system is same as that of Zhang et al. (2008).
1.3 Simulation of coupled multiple processes
Gas flow within coal seams differs significantly from that of conventional reservoirs.
Detailed studies have examined the storage and transport mechanisms of gas in coal seams.
In situ and laboratory data indicate that the storage and flow of gas in coal seams is
associated with the matrix structure of coal and the absorption or desorption of gas. Coal is
a naturally fractured dual-porosity reservoir, consisting of micro-porous matrix and cleats.
Most of the gas is initially stored within micro-pores in the absorbed state. When gas
recovery begins, the gas desorbs and diffuses from the matrix to the cleats due to the
concentration gradient. The rate of gas flow through the cleats is considered to be
controlled by the permeability of the coal seam. Gas flow within coal seams is a complex
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-15
physical and chemical process coupling solid deformation, gas desorption and gas
movement. The complexities of process interactions exert a strong control on ultimate
behaviour—these include linear physical interactions, but also the development of material
nonlinearities that irreversibly alter the affected media.
According to Minkoff et al. (2003), there are three basic algorithms for the simulation of
coupled processes: one-way coupling, loose coupling, and full coupling. For one-way
coupling, separate sets of equations are solved independently over the same total time
interval. Periodically, output from one simulator is passed as input to the other; however,
information is passed in only one direction. A loose coupling resides somewhere between
full and one-way coupling. In loose coupling, different sets of equations are solved
independently (as in one-way coupling), but information is passed at designated time
intervals in both directions between the simulators. For a full coupling, a single set of
equations (generally a large system of non-linear coupled partial differential equations)
incorporating all of the relevant physics needs to be solved simultaneously.
1.3.1 One-way coupling
Coal porosity and permeability models have a variety of forms when specific conditions are
imposed. When the change in total stress is equal to zero, 0 , both coal porosity and
permeability are independent of the total stress. Similarly, when the coal sample is under
the uniaxial strain condition and the overburden load remains unchanged, they are also
independent of the total stress. In this review, studies under these assumptions are
considered as one-way coupling.
Balla (1989) developed a mathematical modelling to simulate methane flow in a borehole
coal mining system, which considered both the sorption phenomenon of methane and, as a
consequence of this, a change in the permeability of the coal. Young (1998) used the
nonequilibrium and pseudosteady state formulations to simulate coalbed methane
production performance, in which the diffusion coefficient is considered to be dependent
on the geometry of the matrix elements and time. The stress-induced changes in cleat
porosity and permeability were included, and the matrix shrinkage due to release of
adsorbed gas are also considered. Similarly, Gilman and Beckie (2000) proposed a
simplified mathematical model of methane movement in a coal seam taking into account
the following features: a relatively regular cleat system, adsorptive methane storage, an
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-16
extremely slow mechanism of methane release from the coal matrix into cleats and a
significant change of permeability due to desorption.
Considering coal to be a triple-porosity system, the implementation of a bidisperse pore-
diffusion model in a coalbed reservoir simulator was discussed by Shi and Durucan (2005),
in which the gas adsorption is assumed to take place only in the micropores with
macropores providing storage for free gas, as well as tortuous paths for gas transport
between the micropores and cleats. Recently, Ross et al. (2009) presented a 3D stochastic
reservoir model to address gas buoyancy and leakage associated with CO2 injection in
coalbeds by using geostatistical techniques and history-matching. More recently, a
mathematical model was developed by Ozdemir (2009) to predict coal bed methane (CBM)
production and carbon dioxide (CO2) sequestration in a coal seam accounting for the coal
seam properties. It was assumed that the flow in a coal seam is a two-phase flow including
a water phase and a gas phase governed by Darcy's law while the flow in the coal matrix is
a diffusional flow governed by Fick's Law, but constant absolute permeability was used in
this study.
These prior studies did not accommodate geomechanical influences related to the role of
changes in total stress on performance. Zhao and Valliappan (1995) derived the governing
equations of methane gas migration in coal seams, which considered the effect of
deformation of the medium, two-phase flow and mass/gas transfer on methane transport
processes in porous media. The permeability magnitudes for both gas and water flow were
considered to be both the same and constant. This work was extended by Valliappan and
Wohua (1996), who presented the development of a mathematical model for methane gas
migration in coal seams, mainly focusing on the coupling between the gas flow and
deformation of solid coal. Anisotropic flow and the effect of diffusion of adsorbed
methane has been considered in this study, but assumed that the porosity of the coal seam
does not change when the gas pressure varies and was constant within an individual
element.
1.3.2 Loose coupling
When coal is under variable stress conditions and the impact of coal fractures and gas
compositions is considered, coal porosity and permeability models are defined as a function
of effective stress, coal matrix-fracture interactions, and gas compositions. Under these
conditions, important non-linear responses due to the effective stress effects need to be
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-17
recovered. This can be achieved through loose coupling. In loose coupling, different sets of
equations are solved independently (as in one-way coupling), but information is passed at
designated time intervals in both directions between the simulators.
The theory describing fluid-solid coupling was first presented by Biot (1941) for a linear
solid where deformation occurred but where no updating was applied to changes in
permeabilities due to infinitesimal changes in porosity. The original Biot theory is for a
single-fluid/single-solid model consistent with single porosity behaviour. Naturally
fractured reservoirs are often modelled by the dual-porosity (overlapping continua) type of
concept developed by Barenblatt et al. (1960). Models incorporating both Biot
poroelasticity and Barenblatt dual-porosity concepts have been studied by many authors
(Duguid and Lee, 1977; Valliappan and Khalili-Naghadeh, 1990; Chen and Teufel, 1997). A
mathematical model of coupled solid-gas for gas flow in coal seams is presented by Zhao et
al. (1994), but the permeability was considered to be constant and the sorption-induced
strain was not coupled in this study. Zhao et al. (2004) extended their work by considered
permeability is a function of volumetric stress and pore pressure to emphasize the coupled
interaction laws between solid deformations and gas seepage within the coal matrix and
fractures, but the influence of sorption-induced strain on permeability change was still not
considered. Gu and Chalaturnyk (2005, 2006) utilized the dynamic change of permeability
for geomechanical and reservoir explicit-sequential coupling simulations, where the
geomechanical simulation is implemented for generalized deformation and stress change
predictions, while multiphase flow is simulated with an appropriate reservoir simulator.
Recently, Gu and Chalaturnyk (2010) established new porosity and permeability models
used for reservoir and geomechanical coupled simulation, which considered a
discontinuous coal mass (containing cleats and matrix) as an equivalent continuum elastic
medium and the anisotropic permeability of coalbeds. Matrix shrinkage/swelling due to gas
desorption/adsorption, thermal expansion due to temperature change, and mechanical
parameters, are included in their work. Similar work has also been conducted by Wang et al.
(2010). Connell (2009) conducted a coupled numerical model and used it to investigate the
applicability of these geomechanical assumptions for gas drainage from coal seams. The
modelling approach involved coupling the existing coal seam gas reservoir simulator,
SIMED II, with the geomechanical simulator, FLAC3D. While SIMED II was used to
simulate gas migration in a hypothetical coal seam and a series of production scenarios,
FLAC3D simulated the geomechanical response of the coal and the adjacent non-coal
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-18
geological formations to fluid pressure and gas content changes imported from SIMED II.
Recently, this work was extended to CO2-ECBM (Connell and Detournay, 2009). Similarly,
Zuo-Tang et al. (2009) proposed a deformation-flow coupled model to address CO2-
geosequestration enhanced coal bed methane recovery. The permeability is considered to
be a function of effective stress, but the influence of the sorption-induced strain on
permeability was not coupled. The interaction between mechanical deformation and fluid
flow in fault zones was addressed by Cappa and Rutqvist (2010), and the TOUGH–FLAC
simulator was applied to supercritical CO2 injection, geomechanics, and ground surface
deformations. Liu et al. (2010c) performed a coupled reactive flow and transport modelling
to simulate large scale CO2 injection. The governing mathematical equation employed in
TOUGHREACT to describe geochemical processes involving fluid-rock interactions.
More Recently, considering the Klinkenberg effect, Wei and Zhang (2010) developed a
two-dimensional, two-phase, triple-porosity/dual-permeability, coupled fluid-flow and
geomechanics CBM simulator for modelling gas and water production, and the coupling
effects of effective stress and micro-pore swelling/shrinkage are modelled with the coupled
fluid-flow and geomechanical deformation approach.
1.3.3 Full coupling
In order to recover important non-linear responses due to the effective stress effects,
mechanical influences need to be rigorously coupled with the gas transport system. This
can be achieved through the full coupling approach. For full coupling, a single set of
equations (generally a large system of non-linear coupled partial differential equations)
incorporating all of the relevant physics will be solved simultaneously.
A coupled mathematical model for solid deformation and gas flow is proposed and is
implemented by Zhu et al. (2007). The finite element method was used to solve the
coupled processes together with Klinkenberg effect. The empirical permeability expression
obtained by Harpalani and Schraufnagel (1990) was used. Similarly, Zhang et al. (2008)
conducted another study on coupled gas flow and coal deformation processes
incorporating the newly developed permeability model, which considers the controlling
factors of the volume occupied by the free-phase gas, the volume occupied by the
adsorbed phase gas, the coal mechanical deformation induced pore volume change, and the
sorption induced coal pore volume change. Based on Zhang et al.'s work, equivalent
poroelastic models (Liu et al., 2010a,b) were developed to simulate the interactions of
multiple processes triggered by the injection or production of single gas. Chen et al. (2009,
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-19
2010a) extended these single poroelastic models to include the flow and transport of gas
mixtures (binary gasses: CO2 and CH4). Wu et al. (2010b, 2011) extended these models
further to a dual poroelastic model (dual solid media- coal matrix and fracture) for both
single gas and binary gas systems. Based on the variable saturation model, Liu and Smirnov
(2008) solved a set of related variables regarding CO2 sequestration in coalbeds, including
capillary pressure, relative permeability, porosity, coupled adsorption model, concentration
and temperature equations. With the same assumptions, the above work was extended to
address the importance of structural deformation effects on carbon sequestration
modelling, which affects the fluid flow and leads to a faster drop of the resulting capillary
pressure and relative permeability of the gas phase (Liu and Smirnov, 2009).
In summary, to define a fully coupled computer simulator for the full mechanics of coal-
gas interactions, a single set of equations (generally a large system of non-linear coupled
partial differential equations) incorporating all of the relevant physics need to be derived.
Full coupling is often the preferred method for simulating multiple types of physics
simultaneously since it should theoretically produce the most realistic results. This could
be the best approach to represent important non-linear responses due to the effective
stress effects when mechanical influences are rigorously coupled with the gas transport
system.
1.4 Research goals
This study was undertaken with the aims of:
i. Carrying out fundamental to gain understanding of the coal properties response,
coal permeability in particular, to variations in effective stress, temperature and gas
pressure and types.
ii. Developing a permeability model to define the evolution of gas sorption-induced
permeability anisotropy.
iii. Conducting a series of gas flow-through experiments to investigate the sole
influence of effective stress and sorption-induced strain on permeability change.
iv. Developing a phenomenological permeability model to quantify coal permeability
evolution under the unconstrained conditions of variable stress by combining the
effect of swelling strain with that of the mechanical effective stress.
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-20
v. Studying the role of heterogeneity on permeability evolution under the
unconstrained swelling conditions.
vi. Conducting a finite element (FE) numerical model with the dual-porosity/dual-
permeability geometry to quantify the impact of various parameters on the CBM
production.
1.5 Thesis outline
This thesis comprises eight chapters. The seven chapters following this introductory
chapter are arranged as follows:
Fundamental experiments is conducted in Chapter 2 to gain understanding of the coal
properties response, coal permeability in particular, to variations in effective stress,
temperature, gas pressure and types. Chapter 3 presents a permeability model to define the
evolution of gas sorption-induced permeability anisotropy under the full spectrum of
mechanical conditions spanning prescribed in-situ stresses through constrained
displacement.
A series of gas flow-through experiments were conducted to quantify the sole influence of
the effective stress and sorption-induced strain on permeability evolution in Chapter 4, and
based on the observations, a phenomenological permeability model is developed in
Chapter 5 to explain this enigmatic behavior of coal permeability evolution under the
influence of gas sorption by combining the effect of swelling strain with that of the
mechanical effective stress.
In Chapter 6, the role of heterogeneity on permeability evolution is numerically conducted
under the unconstrained swelling conditions with the heterogeneous distributions of
Young's modulus and Langmuir strain constant in the vincity of the fracture. In Chapter 7,
a dual-porosity/dual-permeability system is incorporated to quantify the impact of various
parameters on the CBM production, particularly the sorption time and shape factor.
Finally, Chapter 8 summarizes the main outcomes of this research, along with suggestions
for future studies.
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-21
1.6 References
Al-Yousef, H.Y., 2005. Permeability Anisotropy Measurement on Whole Cores- Analytical Solution and
Application, SPE Middle East Oil and Gas Show and Conference. Society of Petroleum Engineers,
Kingdom of Bahrain.
Bai, M., Elsworth, D., Inyang, H.I. and Roegiers, J.C., 1997. Modeling Contaminant Migration with Linear
Sorption in Strongly Heterogeneous Media. Journal of Environmental Engineering, 123(11): 1116-
1125.
Bai, M., Elsworth, D. and Roegiers, J.-C., 1993. Multiporosity/multipermeability approach to the simulation
of naturally fractured reservoirs. Water Resour. Res., 29(6): 1621-1633.
Bai, M., Roegiers, J.-C. and Elsworth, D., 1995. Poromechanical response of fractured-porous rock masses.
Journal of Petroleum Science and Engineering, 13(3-4): 155-168.
Balla, L., 1989. Mathematical modeling of methane flow in a borehole coal mining system. Transport in
Porous Media, 4(2): 199-212.
Barenblatt, G.I., Zheltov, I.P. and Kochina, I.N., 1960. Basic concepts in the theory of seepage of
homogeneous liquids in fissured rocks. J. Appl. Math. Mech., 24: 1286-1303.
Biot, M.A., 1941. General Theory of Three-Dimensional Consolidation. Journal of Applied Physics, 12(2):
155-164.
Cappa, F. and Rutqvist, J., 2010. Modelling of coupled deformation and permeability evolution during fault
reactivation induced by deep underground injection of CO2. International Journal of Greenhouse
Gas Control., doi:10.1016/j.ijggc.2010.08.005.
Chen, H.-Y. and Teufel, L.W., 1997. Coupling Fluid-Flow and Geomechanics in Dual-Porosity Modeling of
Naturally Fractured Reservoirs, SPE Annual Technical Conference and Exhibition. Society of
Petroleum Engineers, Inc., San Antonio, Texas.
Chen, Z., Liu, J., Elsworth, D., Connell, L. and Pan, Z., 2009. Investigation of CO2 Injection Induced Coal-
Gas Interactions, 43rd U.S. Rock Mechanics Symposium & 4th U.S. - Canada Rock Mechanics
Symposium. American Rock Mechanics Association, Asheville, North Carolina.
Chen, Z., Liu, J., Connell, L., Pan, Z. and Zhou, L., 2008. Impact of Effective Stress and CH4-CO2 Counter-
Diffusion on CO2 Enhanced Coalbed Methane Recovery, SPE Asia Pacific Oil and Gas Conference
and Exhibition, Perth, Australia.
Chen, Z., Liu, J., Elsworth, D., Connell, L.D. and Pan, Z., 2010a. Impact of CO2 injection and differential
deformation on CO2 injectivity under in-situ stress conditions. International Journal of Coal Geology,
81(2): 97-108.
Chen, Z., Liu, J., Pan, Z., Connell, L. and Elsworth, D., 2010b. Relations Between Coal Permeability And
Directional Strains And Their Application to San Juan Basin, 44th U.S. Rock Mechanics Symposium
and 5th U.S.-Canada Rock Mechanics Symposium. American Rock Mechanics Association, Salt Lake
City, Utah.
Clarkson, C.R., 2008. Case Study: Production Data and Pressure Transient Analysis of Horseshoe Canyon
CBM Wells, CIPC/SPE Gas Technology Symposium 2008 Joint Conference. Society of Petroleum
Engineers, Calgary, Alberta, Canada.
Connell, L.D., 2009. Coupled flow and geomechanical processes during gas production from coal seams.
International Journal of Coal Geology, 79(1-2): 18-28.
Connell, L.D. and Detournay, C., 2009. Coupled flow and geomechanical processes during enhanced coal
seam methane recovery through CO2 sequestration. International Journal of Coal Geology, 77(1-2):
222-233.
Connell, L.D., Lu, M. and Pan, Z., 2010. An analytical coal permeability model for tri-axial strain and stress
conditions. International Journal of Coal Geology, 84(2): 103-114.
Cui, X. and Bustin, R.M., 2005. Volumetric strain associated with methane desorption and its impact on
coalbed gas production from deep coal seams. AAPG Bulletin, 89(9): 1181-1202.
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-22
Douglas Jr, J., Hensley, J.L. and Arbogast, T., 1991. A dual-porosity model for waterflooding in naturally
fractured reservoirs. Computer Methods in Applied Mechanics and Engineering, 87(2-3): 157-174.
Duguid, J.O. and Lee, P.C.Y., 1977. Flow in fractured porous media. Water Resour. Res., 13(3): 558-566.
Elsworth, D. and Bai, M., 1992. Flow-Deformation Response of Dual-Porosity Media. Journal of
Geotechnical Engineering, 118(1): 107-124.
Gash, B.W., Richard, F.V., Potter, G. and Corgan, J.M., 1992. The effects of cleat orientation and confining
pressure on cleat porosity, permeability and relative permeability in coal, SPWLA/SCA Symposium,
Oklahoma City.
Gilman, A. and Beckie, R., 2000. Flow of Coal-Bed Methane to a Gallery. Transport in Porous Media, 41(1):
1-16.
Gray, I., 1987. Reservoir Engineering in Coal Seams: Part 1-The Physical Process of Gas Storage and
Movement in Coal Seams. SPE Reservoir Engineering, 2(1): 28-34.
Gu, F. and Chalaturnyk, J.J., 2005. Analysis of Coalbed Methane Production by Reservoir and
Geomechanical Coupling Simulation. 44(10).
Gu, F. and Chalaturnyk, R., 2010. Permeability and porosity models considering anisotropy and
discontinuity of coalbeds and application in coupled simulation. Journal of Petroleum Science and
Engineering, 74(3-4): 113-131.
Gu, F., Chalaturnyk, R.J., 2006. Numerical Simulation of Stress and Strain Due to Gas Sorption/Desorption
and Their Effects on In Situ Permeability of Coalbeds. Journal of Canadian Petroleum Technology
45 (10), 52-62.
Harpalani, S. and Chen, G., 1997. Influence of gas production induced volumetric strain on permeability of
coal. Geotechnical and Geological Engineering, 15(4): 303-325.
Harpalani, S. and Schraufnagel, R.A., 1990. Shrinkage of coal matrix with release of gas and its impact on
permeability of coal. Fuel, 69(5): 551-556.
Izadi, G., Wang, S., Elsworth, D., Liu, J., Wu, Y., Pone, D., 2011. Permeability evolution of fluid-infiltrated
coal containing discrete fractures. International Journal of Coal Geology 8, 202-211.
Koenig, R.A. and Stubbs, P.B., 1986. Interference Testing of a Coalbed Methane Reservoir, SPE
Unconventional Gas Technology Symposium. Society of Petroleum Engineers, Inc., Louisville,
Kentucky.
Kuuskraa, V.A., Boyer, C.M. and Kelafant, J.A., 1992. Hunt for quality basins goes abroad. Oil & Gas
Journal, 5: 72-77.
Liu, F., Lu, P., Zhu, C. and Xiao, Y., 2010c. Coupled reactive flow and transport modeling of CO2
sequestration in the Mt. Simon sandstone formation, Midwest U.S.A. International Journal of
Greenhouse Gas Control, doi:10.1016/j.ijggc.2010.08.008.
Liu, G. and Smirnov, A.V., 2008. Modeling of carbon sequestration in coal-beds: A variable saturated
simulation. Energy Conversion and Management, 49(10): 2849-2858.
Liu, G. and Smirnov, A.V., 2009. Carbon sequestration in coal-beds with structural deformation effects.
Energy Conversion and Management, 50(6): 1586-1594.
Liu, H.-H. and Rutqvist, J., 2010. A New Coal-Permeability Model: Internal Swelling Stress and Fracture–
Matrix Interaction. Transport in Porous Media, 82(1): 157-171.
Liu, J., Chen, Z., Elsworth, D., Miao, X. and Mao, X., 2010a. Evaluation of stress-controlled coal swelling
processes. International Journal of Coal Geology, 83(4): 446-455.
Liu, J., Chen, Z., Elsworth, D., Miao, X. and Mao, X., 2010b. Linking gas-sorption induced changes in coal
permeability to directional strains through a modulus reduction ratio. International Journal of Coal
Geology, 83(1): 21-30.
Liu, J., Chen, Z. and Qu, H., 2010c. Multiphysics of Coal-Gas Interactions: The Scientific Foundation for
CBM Extraction, SPE Asia Pacific Oil and Gas Conference and Exhibition, Brisbane, Queensland,
Australia.
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-23
Liu, J., Chen, Z., Elsworth, D., Miao, X., Mao, X., 2011. Evolution of coal permeability from stress-
controlled to displacement-controlled swelling conditions. Fuel 90, 2987-2997.
Liu, J. and Elsworth, D., 1997. Three-dimensional effects of hydraulic conductivity enhancement and
desaturation around mined panels. International Journal of Rock Mechanics and Mining Sciences,
34(8): 1139-1152.
Lu, M. and Connell, L.D., 2007. A model for the flow of gas mixtures in adsorption dominated dual porosity
reservoirs incorporating multi-component matrix diffusion: Part I. Theoretical development. Journal
of Petroleum Science and Engineering, 59(1-2): 17-26.
Ma, Q., Harpalani, S. and Liu, S., 2011. A simplified permeability model for coalbed methane reservoirs
based on matchstick strain and constant volume theory. International Journal of Coal Geology, 85
(1), 43-48.
Massarotto, P., Golding, S.D. and Rudolph, V., 2009. Constant volume CBM reservoirs: an important
principle, International Coalbed Methane Symposium, Tuscaloosa, Alabama.
Minkoff, S.E., Stone, C.M., Bryant, S., Peszynska, M. and Wheeler, M.F., 2003. Coupled fluid flow and
geomechanical deformation modeling. Journal of Petroleum Science and Engineering, 38(1-2): 37-56.
Murray, D.K., 1996. Coalbed methane in the USA: analogues for worldwide development. Geological
Society, London, Special Publications, 109(1): 1-12.
Nowacki, W., 1995. Dynamic Problems of Thermoelasticity. Noordhoff, Leiden, The Netherlands.
Ouyang, Z. and Elsworth, D., 1993. Evaluation of groundwater flow into mined panels. International
Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts, 30(2): 71-79.
Ozdemir, E., 2009. Modeling of coal bed methane (CBM) production and CO2 sequestration in coal seams.
International Journal of Coal Geology, 77(1-2): 145-152.
Palmer, I., 2008. Coalbed methane wells are cheap, but permeability can be expensive! Energy Tribune: 10-
13.
Palmer, I. and Mansoori, J., 1996. How Permeability Depends on Stress and Pore Pressure in Coalbeds: A
New Model, SPE Annual Technical Conference and Exhibition. Society of Petroleum Engineers,
Inc., Denver, Colorado.
Palmer, I.D., Mavor, M. and Gunter, B., 2007. Permeability changes in coal seams during production and
injection, International Coalbed Methane Symposium, University of Alabama, Tuscaloosa, Alabama.
Paper 0713.
Pan, Z. and Connell, L.D., 2007. A theoretical model for gas adsorption-induced coal swelling. International
Journal of Coal Geology, 69(4): 243-252.
Pan, Z., Connell, L.D. and Camilleri, M., 2010. Laboratory characterisation of coal reservoir permeability for
primary and enhanced coalbed methane recovery. International Journal of Coal Geology, 82(3-4):
252-261.
Pan, Z. and Connell, L.D., 2011. Modelling of anisotropic coal swelling and its impact on permeability
behaviour for primary and enhanced coalbed methane recovery. International Journal of Coal
Geology, 85 (3-4), 257-267.
Patching, T.H., 1965. Variations in permeability of coal, Proc. 3rd Rock Mech. Symp, Univ. of Toronto.
U.K., pp. 185-199.
Patching, T.H., 1970. Retention and release of gas in coal- a review. Canadian Mining and Metallurgical
Bulletin, 63(703): 1302-1308.
Pekot, L.J. and Reeves, S.R., 2002. Modeling the effects of matrix shrinkage and differential swelling on
coalbed methane recovery and carbon sequestration, U.S. Department of Energy, DE-FC26-
00NT40924.
Pini, R., Ottiger, S., Burlini, L., Storti, G. and Mazzotti, M., 2009. Role of adsorption and swelling on the
dynamics of gas injection in coal. J. Geophys. Res., 114(B4): B04203.
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-24
Pomeroy, C.D. and Robinson, D.J., 1967. The effect of applied stresses on the permeability of a middle rank
coal to water. International Journal of Rock Mechanics and Mining Sciences & Geomechanics
Abstracts, 4(3): 329-343.
Reiss, L.H., 1980. The Reservoir Engineering Aspects of Fractured Formations. Gulf Publishing Co.,
Houston.
Robertson, E.P., 2005. Modeling Permeability in Coal Using Sorption-Induced Strain Data, SPE Annual
Technical Conference and Exhibition. Society of Petroleum Engineers, Dallas, Texas.
Robertson, E.P. and Christiansen, R.L., 2006. A Permeability Model for Coal and Other Fractured, Sorptive-
Elastic Media, SPE Eastern Regional Meeting. Society of Petroleum Engineers, Canton, Ohio, USA.
Ross, H.E., Hagin, P. and Zoback, M.D., 2009. CO2 storage and enhanced coalbed methane recovery:
Reservoir characterization and fluid flow simulations of the Big George coal, Powder River Basin,
Wyoming, USA. International Journal of Greenhouse Gas Control, 3(6): 773-786.
Seidle, J.P., Jeansonne, M.W. and Erickson, D.J., 1992. Application of Matchstick Geometry To Stress
Dependent Permeability in Coals, SPE Rocky Mountain Regional Meeting, Casper, Wyoming.
Seidle, J.R. and Huitt, L.G., 1995. Experimental Measurement of Coal Matrix Shrinkage Due to Gas
Desorption and Implications for Cleat Permeability Increases, International Meeting on Petroleum
Engineering. Society of Petroleum Engineers, Inc., Beijing, China.
Shi, J.-Q. and Durucan, S., 2005. Gas Storage and Flow in Coalbed Reservoirs: Implementation of a
Bidisperse Pore Model for Gas Diffusion in Coal Matrix. SPE Reservoir Evaluation & Engineering,
8(2): 169-175.
Shi, J.Q. and Durucan, S., 2004. Drawdown Induced Changes in Permeability of Coalbeds: A New
Interpretation of the Reservoir Response to Primary Recovery. Transport in Porous Media, 56(1): 1-
16.
Siriwardane, H. et al., 2009. Influence of carbon dioxide on coal permeability determined by pressure
transient methods. International Journal of Coal Geology, 77(1-2): 109-118.
Somerton, W.H., Söylemezoglu, I.M. and Dudley, R.C., 1975. Effect of stress on permeability of coal.
International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts, 12(5-6):
129-145.
Sparks, D.P., McLendon, T.H., Saulsberry, J.L. and Lambert, S.W., 1995. The Effects of Stress on Coalbed
Reservoir Performance, Black Warrior Basin, U.S.A, SPE Annual Technical Conference and
Exhibition. Society of Petroleum Engineers, Inc., Dallas, Texas.
Valliappan, S. and Khalili-Naghadeh, N., 1990. Flow through fissured porous media with deformable matrix.
International Journal for Numerical Methods in Engineering, 29(5): 1079-1094.
Valliappan, S. and Wohua, Z., 1996. Numerical modelling of methane gas migration in dry coal seams.
International Journal for Numerical and Analytical Methods in Geomechanics, 20(8): 571-593.
van Bergen, F., Spiers, C., Floor, G. and Bots, P., 2009. Strain development in unconfined coals exposed to
CO2, CH4 and Ar: Effect of moisture. International Journal of Coal Geology, 77(1-2): 43-53.
Wang, G.X., Massarotto, P. and Rudolph, V., 2009. An improved permeability model of coal for coalbed
methane recovery and CO2 geosequestration. International Journal of Coal Geology, 77(1-2): 127-
136.
Wang, G.X., Wei, X.R., Wang, K., Massarotto, P. and Rudolph, V., 2010. Sorption-induced
swelling/shrinkage and permeability of coal under stressed adsorption/desorption conditions.
International Journal of Coal Geology, 83(1): 46-54.
Wang, S., Elsworth, D. and Liu, J., 2011. Permeability evolution in fractured coal: The roles of fracture
geometry and water-content. International Journal of Coal Geology, doi:10.1016/j.coal.2011.04.009.
Warren, J.E. and Root, P.J., 1963. The Behavior of Naturally Fractured Reservoirs. 3(3): 245 - 255.
Wei, Z. and Zhang, D., 2010. Coupled fluid-flow and geomechanics for triple-porosity/dual-permeability
modeling of coalbed methane recovery. International Journal of Rock Mechanics and Mining
Sciences, 47 (8), 1242-1253.
School of Mechanical and Chemical Engineering
Chapter 1 The University of Western Australia
1-25
White, C.M. et al., 2005. Sequestration of Carbon Dioxide in Coal with Enhanced Coalbed Methane
RecoveryA Review†. Energy & Fuels, 19(3): 659-724.
Wong, R.C., 2003. A model for strain-induced permeability anisotropy in deformable granular media.
Canadian Geotechnical Journal, 40: 95-106.
Wu, Y. et al., 2010b. Dual poroelastic response of a coal seam to CO2 injection. International Journal of
Greenhouse Gas Control, 4(4): 668-678.
Wu, Y., Liu, J., Elsworth, D., Miao, X. and Mao, X., 2010a. Development of anisotropic permeability during
coalbed methane production. Journal of Natural Gas Science and Engineering, 2(4): 197-210.
Wu, Y., Liu, J., Chen, Z., Elsworth, D., Pone, D., 2011. A dual poroelastic model for CO2-enhanced
coalbed methane recovery. International Journal of Coal Geology 86, 177-189.
Young, G.B.C., 1998. Computer modeling and simulation of coalbed methane resources. International
Journal of Coal Geology, 35(1-4): 369-379.
Zhang, H., Liu, J. and Elsworth, D., 2008. How sorption-induced matrix deformation affects gas flow in
coal seams: A new FE model. International Journal of Rock Mechanics and Mining Sciences, 45(8):
1226-1236.
Zhao, C. and Valliappan, S., 1995. Finite element modelling of methane gas migration in coal seams.
Computers & Structures, 55(4): 625-629.
Zhao, Y., Hu, Y., Zhao, B. and Yang, D., 2004. Nonlinear Coupled Mathematical Model for Solid
Deformation and Gas Seepage in Fractured Media. Transport in Porous Media, 55(2): 119-136.
Zhao, Y., Jin, Z. and Sun, J., 1994. Mathematical model for coupled solid deformation and methane flow in
coal seams. Applied Mathematical Modelling, 18(6): 328-333.
Zhu, W.C., Liu, J., Sheng, J.C. and Elsworth, D., 2007. Analysis of coupled gas flow and deformation
process with desorption and Klinkenberg effects in coal seams. International Journal of Rock
Mechanics and Mining Sciences, 44(7): 971-980.
Zuo-tang, W. et al., 2009. Simulation of CO2-geosequestration enhanced coal bed methane recovery with a
deformation-flow coupled model. Procedia Earth and Planetary Science, 1(1): 81-89.
School of Mechanical and Chemical Engineering
Chapter 2 The University of Western Australia
2–1
Chapter 2
Laboratory Characterisation of Fluid Flow in Coal with respect to Gas Type and Temperature
Zhejun Pana, Zhongwei Chenb, Luke D. Connella, and Nicolas Luptona
a CSIRO Earth Science and Resource Engineering, Private Bag 10, Clayton South 3169, Australia
b School of Mechanical and Chemical Engineering, The University of Western Australia, WA 6009, Australia
Abstract: Coalbed methane is an important unconventional gas resource. However, a
significant amount of the gas resource can not be produced through pressure drawdown
due to the nature of gas storage through adsorption. Enhanced coalbed methane via
injecting gasses such as carbon dioxide and nitrogen is regarded as a viable option to
increase the recovery of the reservoir methane. Injecting CO2 in deep, unminable coal
seams is also considered an option for CO2 sequestration to reduce greenhouse gas
emissions. Often the gas is injected at a different temperature to the seam, for instance, a
hot flue gas injection to the coal seam to enhance coalbed methane recovery. Thus a locally
temperature affected zone will form around the injection well. Although the affect of
temperature on gas adsorption in coal is well understood, how temperature affects the gas
effective permeability for different gasses and the overall gas flow behaviour has not been
extensively studied. In this work, laboratory measurements are carried out on an Australian
coal sample using Helium, N2, CH4 and CO2 at two temperatures, 35 ºC and 40 ºC. Gas
adsorption isotherms, sorption-induced coal swelling isotherms, effective gas permeability
under tri-axial conditions with respect to pore and confining pressures and geomechanical
properties are measured. The experimental results show that the amount of gas adsorbed at
35 ºC is slightly higher than that at 40 ºC for the same pressure for all three adsorptive
gasses, N2, CH4 and CO2. This is consistent with other experimental studies presented in
the literature. Adsorption induced coal swelling for the same pressures is also slightly
higher at 35 ºC and is in accord with gas adsorption behaviour. When the swelling strain is
plotted with respect to the quantity adsorbed, the measurements for the various gasses at
different temperatures tend to fall onto a single trendline, which is almost linear.
Geomechanical properties, which are different for each gas, however, do not show a
significant difference between the two temperatures. Nevertheless, the effective gas
School of Mechanical and Chemical Engineering
Chapter 2 The University of Western Australia
2–2
permeability tends to be higher at higher temperature for all gasses. More importantly, cleat
compressibility, a key property in determining coal reservoir permeability behaviour, is the
least when measured with N2 and the largest using CO2 among the three gasses. This may
suggest that the cleat compressibility is correlated with adsorption and adsorption-induced
swelling. Part of the reason for this may be attributed to the impact of adsorption induced
coal swelling on the cleat porosity. In summary, these measurements show that gas species
and temperature may have a significant impact on gas flow behaviour in coal. Hence,
injected gas composition and temperature should be considered as parameters in the CO2/N2
injection strategy to optimize the ECBM process.
Keywords: coalbed methane; enhanced coalbed methane; cleat compressibility;
permeability; swelling; adsorption
School of Mechanical and Chemical Engineering
Chapter 2 The University of Western Australia
2–3
2.1 Introduction
Coalbed methane (CBM) or coal seam gas (CSG) is an important unconventional gas
resource. CBM production is mainly by reservoir pressure drawdown and CBM reservoir
permeability is considered the most important reservoir property for CBM production
(Palmer, 2009; Sparks et al., 1995). However, a substantial amount of methane is expected
to be left behind by the primary recovery because reduction of reservoir pressure to lower
than 150 psia (~1 MPa) is generally not considered as practical or economic (Puri and Lee,
1990). Hence, gasses, such as N2 and CO2, can be injected as a strategy to enhance coalbed
methane recovery. Furthermore, Coal is able to adsorb more CO2 than methane at the
same pressure thus it is also considered a viable option to reduce greenhouse gas emissions.
Coal permeability is sensitive to stress (Seidle et al., 1992). Moreover, a unique
characteristic for coal reservoirs is that the coal matrix shrinks as gas desorbs from the coal
and swells as gas adsorbs into the coal matrix, leading to further permeability change under
reservoir conditions. There have been several ECBM trials so involving CO2 injection
(Fujioka et al, 2010; Reeves and Oudinot, 2005; van Bergen et al., 2006; Wong et al., 2007).
A common aspect of these trials has been that CO2 injectivity has been observed to
decrease due to permeability loss as a result of adsorption induced coal swelling. In some
trials CO2 has been injected as liquid (Wong et al., 2007). However, since liquid CO2 has a
high viscosity, which acts to reduce injection rates, it is heated and then injected as
supercritical CO2 which has a lower viscosity (Fujioka et al., 2010). Furthermore, N2 or flue
gas, a mixture of N2 and CO2, can also be injected to enhance coalbed methane recovery.
For all these injection scenarios the injected gas is likely to be at a contrasting temperature
to the seam. While the affects of temperature on gas adsorption in coal is well understood,
how temperature affects the gas effective permeability and the overall gas flow behaviour
has not been extensively studied.
A number of permeability models have been developed to include both the effective stress
and swelling/shrinkage effects. For instance, one of the widely applied models, the Shi and
Durucan (S-D) model, can be expressed as:
13100
VEPP (2-1)
School of Mechanical and Chemical Engineering
Chapter 2 The University of Western Australia
2–4
where is the effective horizontal stress, 0 is the effective horizontal stress at the initial
reservoir pressure, V is the volumetric swelling/shrinkage strain (Shi and Durucan, 2004).
To relate the permeability with effective stress, the equation below is used:
03
0
fc
ekk (2-2)
where fc is referred to as the cleat volume compressibility with respect to changes in the
effective horizontal stress normal to the cleats (Shi and Durucan, 2004). From the above
equations, the coal geomechanical properties, E and , the swelling ratio, V , and the
cleat compressibility, fc , are important parameters to determine the behaviour of reservoir
permeability.
In this work, laboratory measurements were performed on an Australian coal sample from
the Woodland Hill coal seam, NSW, using Helium, N2, CH4 and CO2 at two temperatures,
35 ºC and 40 ºC. Gas adsorption isotherms, sorption-induced coal swelling isotherms, gas
effective permeability under tri-axial conditions with respect to the pore and confining
pressures and geomechanical properties are measured. Overall gas flow behaviour under
different temperatures is analysed and discussed.
2.2 Experimental
2.2.1 Experiment apparatus description
A triaxial permeability cell was used for the experimental measurement of gas adsorption
and permeability under hydrostatic conditions. Figure shows the schematic of the Triaxial
Multi-Gas Rig used for this work. Radial and axial displacements are measured at each
adsorption step to obtain swelling strain. Four displacement gauges are installed with two
to measure the axial displacement and the other two to measure the radial displacement.
The two radial displacement gauges was installed perpendicularly. The displacement gauges
are not presented in Figure 2-1 to keep this figure concise. Load and displacement tests are
performed at each pore and confining pressure to evaluate the impact of gas species and
applied stress on geomechanical properties. The load change is up to 1MPa for the load
and displacement tests. The core sample, usually 5cm in diameter and 10 to 15 cm in length,
is wrapped with a thin lead foil then a rubber sleeve before it is installed in the cell. The
thin lead foil is to prevent gas diffusion from the core to the confining fluid at high sample
School of Mechanical and Chemical Engineering
Chapter 2 The University of Western Australia
2–5
pressures (Mazumder et al., 2006). The sample cell and other parts of the rig are in a
temperature controlled cabinet to maintain constant temperature during the experiment.
Figure 2-1. Schematic of the triaxial rig.
2.2.2 Adsorption measurement
Prior to gas adsorption, the void volume, voidV , in the cell is determined by injecting
known quantities of helium from a calibrated gas injection pump. Since helium adsorption
is negligible on coal, the void volume can be determined from measured values of
temperature, pressure and the helium volume injected into the cell. This helium void
volume measurement was performed at a range of pressures to investigate the consistency
of the calculated volume. The mass-balance equation, expressed in volumetric terms, is:
cellpump
voidTZ
P
TZ
P
ZT
VPV
1
1
2
2 (2-3)
where V is the volume injected from the gas injection pump, Z is the compressibility
factor of helium, T is the temperature, P is the pressure, subscripts ―cell‖ and ―pump‖ refer
Upstream
CH4
He
CO2
Downstream
CO2 Tube Heater
Upstream Cylinder Downstream
Cylinder Actuator
111.1 122.2
Vaccum
111.1 122.2
Pressure Sensor
Pressure Sensor
Pressure Sensor
Pressure Sensor
Pressure Sensor
Pressure Sensor
Pressure Sensor
Upstream Injection Pump
Confinement Pump
COMPUTER / DAQ
Water Heater
Differential Pressure transducer
School of Mechanical and Chemical Engineering
Chapter 2 The University of Western Australia
2–6
to conditions in the cell and pump, respectively, and subscripts ―1‖ and ―2‖ refer to
conditions in the cell before and after injection of gas from the pump, respectively. This
void volume is used in subsequent measurements of adsorption.
The Gibbs excess adsorption (also known as the excess adsorption) is calculated directly
from the experimental quantities. For pure-gas adsorption measurements, a known
quantity, injn , of gas (e.g., methane) is injected from the gas injection pump into the cell.
Some of the injected gas will be adsorbed, and the remainder, voidn , will exist in the
equilibrium bulk (gas) phase in the cell. A mass balance is used to calculate the amount
adsorbed, Gibbs
adsn , as:
voidinj
Gibbs
ads nnn (2-4)
The amount injected can be determined from pressure, temperature and volume
measurements of the pump:
pump
injZRT
VPn
(2-5)
The amount of gas in the void volume is calculated from conditions at equilibrium in the
cell:
cell
voidvoid
ZRT
PVn
(2-6)
In Equations (2-5) and (2-6) , Z is the compressibility factor of the pure gas at the
corresponding conditions of temperature and pressure. The above steps are repeated
sequentially at higher pressures to yield a complete adsorption isotherm. Equation (2-7) is
used to calculated absolute adsorption from the measured Gibbs excess adsorption:
gasads
adsGibbsads
Absads nn
(2-7)
where ads is the adsorbed phase density, gas is the gas phase density.
Gas compressibility factors and densities for Helium, N2, CH4 and CO2 are calculated from
the NIST webbook at http://webbook.nist.gov/chemistry/fluid/.
2.2.3 Permeability measurement
The transient method of Brace et al. (1968) was used because of the shorter test durations
required compared to steady state measurements. The Brace method involves observing
School of Mechanical and Chemical Engineering
Chapter 2 The University of Western Australia
2–7
the decay of a differential pressure between upstream and downstream vessels across a
sample of interest. This pressure decay is combined with the vessel volumes in the analysis
to relate the flow through the sample and thus determine the permeability (Brace et al.,
1968). The pressure decay curve can be modelled as:
t
du
du ePP
PP
0,0,
(2-8)
where du PP is the pressure difference between the up- and downstream cylinders, in the
experimental facility used for this work, measured by a differential pressure transducer;
0,0, du PP is the pressure difference between the up- and downstream cylinders at initial
stage, t is time and is described below:
du
RVV
VL
k 112
(2-9)
where k is permeability; is the gas compressibility; L is the sample length; RV is the
sample volume; uV and dV are the volume of the up- and downstream cylinders.
2.2.4 Cleat compressibility
To determine cleat compressibility, Seidle et al. (1992) derived a relationship between
permeability and stress by idealising the coal fabric as a collection of matchsticks. This
relationship is Equation (2-2) and was used in combination with the S-D permeability
model as presented in Equation (2-1). In Seidle et al.'s work (1992), water was used to
measure coal permeability and involved a series of permeability measurements at constant
confining pressure but differing pore pressure. In this current work, gasses including
helium, nitrogen, methane and CO2 were used to measure coal permeability. Pan et al.,
(2010) proved that Equation (2-2) derived by Seidle et al. (1992) is valid for the current
permeability measurement using gasses.
Cleat compressibility is defined by:
p
f
f
fP
c
1 (2-10)
where f is cleat porosity and
pP is pore pressure. Hydrostatic stress is defined by (see e.g.
Zimmerman et al., 1986):
pc mPP (2-11)
School of Mechanical and Chemical Engineering
Chapter 2 The University of Western Australia
2–8
where cP is the confining pressure, and m is the effective stress coefficient.
Cleat compressibility can be estimated by fitting Equation (2-2) to permeability
measurements with respect to effective stress.
2.2.5 Swelling measurement
Swelling displacements are measured simultaneously with gas adsorption at a constant
effective stress, which is controlled by tracking the pore pressure. Volume swelling is
approximately represented by:
rarrV 21 (2-12)
where V is the volumetric swelling, 1r and 2r are the two radial strains perpendicular
to each other, ra is the axial strain, which is the average of the results by the two axial
displacement gauge.
2.2.6 Young’s modulus and Poisson’s ratio
Uniaxial stress testing was performed on the coal core. A load was applied in the axial
direction and the axial and radial displacements were monitored and used to calculate the
Young's modulus and Poisson's ratio.
ll
AFE
(2-13)
where F is the load, A is the cross-section area of the core, l is the displacement in the
axial direction, l is the length of the core. Poisson's ratio can be calculated from:
a
r
(2-14)
where r is the radial strain and a is the axial strain. Poisson's ratio can also be calculated
from the relationship between bulk modulus and Young's modulus:
213
EK (2-15)
The effect of the presence of fluid during the load-displacement test to obtain Young's
modulus and Poisson's ratio should be minimal, because the fluid is connected to a big
void volume (about 30 ml), so that the pore pressure change during load-displacement test
is negligible.
School of Mechanical and Chemical Engineering
Chapter 2 The University of Western Australia
2–9
More detailed descriptions on experimental procedures can be found in Pan et al. (2010).
2.3 Results and Discussion
2.3.1 Adsorption isotherm results
Figure 2-2 presents the helium porosity and N2, CH4 and CO2 adsorption results at two
different temperatures on the coal. It can be seen from Figure 2-2(a) that the helium
porosity for this coal is about 9%, which includes the pore volume of the cleat and macro-,
meso- and micro- pores. The adsorption capacity increases from N2 to CH4 to CO2. At
about 8 MPa, the adsorption ratio is about 1:2:3 for the three gasses for this coal. CO2
adsorbs 1.5 times more than CH4, which is slightly lower than many coals with a ratio of
about 2 to 3 (Fitzgerald et al., 2005). CO2 absolute adsorption jumps at 10 MPa, thus not
following the Langmuir curve. The reason for this is not well understood but this
behaviour was also observed by other researchers (e.g. Hall et al., 1994). At two different
temperatures, gasses adsorb slightly more at lower temperature. However, their difference
is minor due to the small temperature difference.
The Langmuir volume and pressure for the three gasses at the measurement temperatures
are summarised in Table 2-1. The Langmuir volumes are slightly larger for N2 and CH4 at
higher temperature. At the same time, the Langmuir pressure is larger as well. The
combined effect makes the Langmuir isotherm lower at higher temperature for the
pressure range considered, reflecting the experimental results. For CO2, the Langmuir
volume is smaller and Langmuir pressure is higher at higher temperature, thus the
Langmuir isotherm is lower at higher temperature. However, it should be noted that the
differences of the isotherm at different temperatures are only marginal and are within
experimental uncertainties.
5.0
6.0
7.0
8.0
9.0
10.0
11.0
0.0 2.0 4.0 6.0 8.0 10.0 12.0Pressure (MPa)
He
liu
m P
oro
sit
y (
%)
increasing pressure (35°C)
Increasing pressure (40°C)
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.0 2.0 4.0 6.0 8.0 10.0 12.0
Pressure (MPa)
Ad
so
rpti
on
(m
mo
l/g)
Pressure up (35°C)
Pressure down (35°C)
Langmuir model (35°C)
Pressure up (40°C)
School of Mechanical and Chemical Engineering
Chapter 2 The University of Western Australia
2–10
(a) Helium porosity (b) N2 adsorption
(c) CH4 adsorption
(d) CO2 adsorption
Figure 2-2. The measured helium porosity and gas adsorption isotherms for the core
sample.
Table 2-1. Langmuir volumes and pressures for the measurements
Gas VL (mmol/g) PL (MPa) VL (mmol/g) PL (MPa)
35 ºC 40 ºC
N2 0.86 5.71 0.90 6.60
CH4 1.25 2.56 1.27 2.78
CO2 1.91 1.40 1.87 1.49
2.3.2 Swelling
The coal strain was measured during the adsorption measurements. For the coal studied,
swelling shows strong anisotropy with swelling strain in the axial direction (perpendicular
to the bedding) almost double that in the radial directions (parallel to the bedding). Figure
2-3 presents an example of the coal swelling strain measurements, which is for CO2 with
respect to pressure and for the two temperatures. It also shows that coal swells slightly
more at lower temperature, which is consistent with the adsorption behaviour with respect
to temperature.
An important characteristic of gas adsorption induced coal swelling is that swelling is
directly related to the adsorbed amount (Pan and Connell, 2007). Figure 2-4 presents the
swelling strain versus adsorbed amount for the three gasses. The results show that the coal
swelling strain versus adsorbed amount for the gasses at the two temperatures almost form
the same trend line, which is close to linear. There are two CO2 points that are outliers,
corresponding to the two adsorption points at higher pressure as shown in Figure 2-2(d).
0.0
0.2
0.4
0.6
0.8
1.0
1.2
0.0 2.0 4.0 6.0 8.0 10.0 12.0Pressure (MPa)
Ad
so
rpti
on
(m
mo
l/g)
Pressure up (35°C)
Pressure down (35°C)
Langmuir model (35°C)
Pressure up (40°C)
0.0
0.5
1.0
1.5
2.0
2.5
0.0 2.0 4.0 6.0 8.0 10.0 12.0
Pressure (MPa)
Ad
so
rpti
on
(m
mo
l/g
)
Pressure up (35°C)
Pressure down (35°C)
Langmuir model (35°C)
Pressure up (40°C)
Not used in Langmuir model fitting
School of Mechanical and Chemical Engineering
Chapter 2 The University of Western Australia
2–11
Figure 2-3. Coal swelling with CO2 adsorption at the two measuring temperatures.
Figure 2-4. Coal swelling for the three gasses as a function of adsorbed amount at the two
measuring temperatures.
2.3.3 Permeability stress behaviour
Permeability was measured at 4 pore pressure steps for each gas up to 10 MPa. Figure 2-5
presents the permeability measurements at 1.0 MPa effective stress at 35 ºC and 40 ºC.
Permeability measurements at each pore pressure step with an effective stress of 2.0 MPa
0.0
0.5
1.0
1.5
2.0
2.5
0 2 4 6 8 10 12Pressure (MPa)
Str
ain
(%
)
Volumetric (40°C)
Volumetric (35°C)
Axial (40°C)
Axial (35°C)
Radial (40°C)
Radial (35°C)
0.0
0.5
1.0
1.5
2.0
2.5
3.0
0 0.5 1 1.5 2 2.5Adsorption amount (mmol/g)
Vo
lum
etr
ic S
we
ll (
%)
N2 (up) 35°C
N2 (down) 35°C
CH4 (up) 35°C
CH4 (down) 35°C
CO2 (up) 35°C
CO2 (down) 35°C
N2 (up) 40°C
CH4 (up) 40°C
CO2 (up) 40°C
School of Mechanical and Chemical Engineering
Chapter 2 The University of Western Australia
2–12
and 3.0 MPa were also performed and will be used to examine the permeability-stress
behaviour and obtain the cleat compressibility in the next section. Permeability decreases
slightly with pore pressure increase for all gasses at the two temperatures. The decrease of
permeability may be caused by (1) Klinkenberg effect, (2) effective stress coefficient,
and/or (3) coal swelling. The permeability at the same pore pressure for different gasses
also show differences, which are also due to the causes outlined above. However, the
decrease is not related to gas adsorption capacity and is somewhat random for this coal. In
our previous work (Pan et al., 2010), permeability was found to decrease in the sequence
from Helium, N2, CH4 to CO2.
Gas permeability measurements at 40 ºC were performed after those at 35 ºC. Helium
permeability at 40 ºC was about 4 times higher than that measured at 35 ºC. This behaviour
is inconsistent with our previous measurement. The cause for this will require further
investigation. Although the comparison of results at different temperatures may be difficult,
the results for each temperature show the same trend.
(a) 35 ºC
(b) 40 ºC
Figure 2-5. Measurements of permeability vs pore pressure at 1 MPa effective stress.
2.3.4 Cleat compressibility
At each pore pressure the permeability was measured with respect to effective stress, by
varying the confining pressure, in order to determine the permeability vs. effective stress
relationship. Figure 2-6 presents the results using N2 at the two measurement temperatures.
Measurements using the other gasses were similar and thus are not plotted. The behaviour
of permeability with respect to effective stress is close to the exponential relationship of
Equation (2-2) and this is used as a fitting function for the measurements to estimate cleat
compressibility.
0
0.1
0.2
0.3
0.4
0.5
0.6
0 2 4 6 8 10 12
Pore Pressure (MPa)
Pe
rme
ab
ilit
y (
md
)
He
N2
CH4
CO20
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 2 4 6 8 10 12
Pore Pressure (MPa)
Perm
eab
ilit
y (
md
)
0
0.5
1
1.5
2
2.5
3
He
liu
m P
erm
ea
bilit
y (
md
)
N2
CH4
CO2
He
School of Mechanical and Chemical Engineering
Chapter 2 The University of Western Australia
2–13
(a) 35 ºC
(b) 40 ºC
Figure 2-6. Measurements of N2 permeablity with respect to effective stress/pressure and
pore pressure.
The calculated cleat compressibilities for each gas at each pore pressure and each
temperature are presented in Figure 2-7. It can be seen from this figure that the cleat
compressibility tends to increase from Helium, N2, CH4 to CO2. This may reflect the
reduction in cleat porosity due to swelling-strain. At lower porosity, cleat compressibility
tends to be higher if the cleat elasticity remains the same. The cleat compressibility results
at 40 ºC are smaller than those in 35 ºC. This may be because of the lower swelling strains
at higher temperature leading to less porosity decrease. However, since the cleat aperture
may be altered by increasing the temperature to 40 ºC, the conclusion of lower
compressibility at higher temperature will require further investigation. These cleat
compressibility results are comparable to the results from Seidle et al.'s work (1992), where
the cleat compressibility is about 0.084MPa-1 for a Warrior Basin coal sample.
(a) Helium
(b) Nitrogen
y = 0.529e-0.198x
R2 = 0.996
y = 0.485e-0.195x
R2 = 0.994
y = 0.438e-0.194x
R2 = 0.991
y = 0.404e-0.211x
R2 = 0.998
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
Effective Pressure (MPa)
N2
eff
ec
tiv
e p
erm
ea
bilit
y (
md
)
0.946 MPa
2.30 MPa
5.23 MPa
10.31 MPa
y = 0.850e-0.116x
R2 = 0.995
y = 0.669e-0.123x
R2 = 0.983
y = 0.607e-0.129x
R2 = 0.988
y = 0.587e-0.152x
R2 = 0.996
0.15
0.25
0.35
0.45
0.55
0.65
0.75
0.85
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5Effective Pressure (MPa)
He
eff
ec
tiv
e p
erm
ea
bilit
y (
md
)
1 MPa
4 MPa
7 MPa
10 MPa
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.10
0 2 4 6 8 10 12Pore Pressure (MPa)
Cle
at
Co
mp
res
sib
ilit
y (
MP
a-1
)
35 °C
40 °C
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0 2 4 6 8 10 12
Pore Pressure (MPa)
Co
mp
res
sib
ilit
y (
MP
a-1
)
35°C
40°C
School of Mechanical and Chemical Engineering
Chapter 2 The University of Western Australia
2–14
(c) Methane
(d) CO2
Figure 2-7. Cleat compressibilities for the various gasses with respect to pore pressure and
temperature.
The large difference of cleat compressibility among different gasses may mean that
permeability changes with different gasses may be significantly different. This is of great
importance to the modelling of ECBM processes. Laboratory measurements are important
to provide the cleat compressibility values to be used in the reservoir simulations.
2.3.5 Geomechanical properties
Load and displacement experiments were performed at each pore pressure and confining
pressure for the gasses and temperature. Figure 2-8 presents the experimental results with
Helium as pore gas. Results with the other gasses are similar and thus not presented, but
Table 2-2 summarises the experimental results for Young's modulus, Poisson's ratio and
bulk modulus for the coal sample with respect to effective stress for each gas and
temperature. The results are averaged with respect to pore pressure.
It can be seen from Table 2-2 that the Young's modulus is smaller at 40 ºC than that
measured at 35 ºC for all gasses. Although the difference is small between the two
temperatures, this may mean that there is a degree of coal softening at higher temperature
and that softening may become important if the temperature change is significant. There is
no apparent trend for the Poisson's ratios between the two temperatures for the different
gasses, but the average among all gasses is slightly higher at lower temperature (0.288 at 35
ºC compared to 0.285 at 40 ºC). The bulk modulus is generally larger at lower temperature.
When averaging among the gasses, the bulk modulus is 1.93 GPa at 35 ºC compared to
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0 2 4 6 8 10 12Pore Pressure (MPa)
Co
mp
ressib
ilit
y (
MP
a-1
)
35°C
40°C
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0 2 4 6 8 10 12
Pore Pressure (MPa)
Co
mp
res
sib
ilit
y (
MP
a-1
)
35°C
40°C
School of Mechanical and Chemical Engineering
Chapter 2 The University of Western Australia
2–15
1.86 GPa at 40 ºC. These results show that temperature has an impact on the coal
properties and the impact may become more important for large temperature differences.
Also from Table 2-2, we can see that the Young's modulus decreases with the increase of
the adsorption capacity among the three adsorbing gasses, N2, CH4 and CO2. The Young's
modulus and bulk modulus are noticeably lower when CO2 is adsorbed to the coal.
Whether adsorption causes coal softening is not clear, since the modulus with Helium
(considered to be non-adsorbing) as the pore gas is not higher than those measured when
adsorbing gasses are used. More work needs to be done to further elucidate the impact of
adsorption on geomechanical properties.
It is quite obvious that the effective stress does have a significant impact on the coal
properties. For instance when helium is the pore gas, the measured Young's modulus at 3.0
MPa effective stress is about 20% larger than that measured at 1.0 MPa effective stress.
Since coal is usually found to be a non-linear elastic rock with load-unload hysteresis, a
larger modulus at higher confining pressure is not surprising and would be important for
reservoir modelling of CBM/ECBM processes.
Figure 2-8. Young's modulus measured with respect to pore pressure at two effective
stresses and temperatures with helium as the pore fluid.
1.0
1.2
1.4
1.6
1.8
2.0
2.2
2.4
2.6
2.8
0.0 2.0 4.0 6.0 8.0 10.0 12.0
Pore Pressure (MPa)
Yo
un
g's
Mo
du
lus
(G
Pa
)
Effective Stress 1.0 MPa (35°C)
Effective Stress 3.0 MPa (35°C)
Effective Stress 1.0 MPa (40°C)
Effective Stress 3.0 MPa (40°C)
School of Mechanical and Chemical Engineering
Chapter 2 The University of Western Australia
2–16
Table 2-2. Averaged geomechanical properties
Young's Modulus (GPa) Poisson's ratio (-) Bulk Modulus (GPa)
Gas Effective
stress (MPa) 35 ºC 40 ºC 35 ºC 40 ºC 35 ºC 40 ºC
He 1.0 2.20 2.10 0.292 0.302 1.77 1.77
3.0 2.60 2.59 0.268 0.291 1.87 2.07
N2 1.0 2.31 2.18 0.300 0.275 1.94 1.61
3.0 2.75 2.68 0.289 0.282 2.17 2.05
CH4 1.0 2.26 2.25 0.291 0.273 1.82 1.65
3.0 2.73 2.70 0.294 0.289 2.21 2.13
CO2 1.0 2.17 2.11 0.279 0.271 1.64 1.54
3.0 2.54 2.50 0.288 0.295 2.00 2.04
2.4 Conclusion
This work presents the experimental measurement of adsorption isotherms, swelling,
permeability, and geomechanical properties for an Australian coal sample in three gasses,
N2, CH4 and CO2 at two temperatures, 35 ºC and 40 ºC. The results show that the gas
adsorption is less at higher temperature, which is consistent with other work reported in
the literature. Measurements of coal swelling strain with respect to adsorbed amount are
close to the same trend line irrespective of gas type and temperature. Permeability
coefficients decline with increased gas pressure. However, the permeability results do not
show a trend in relation to the gas adsorption capacity for this coal. The cleat
compressibility was estimated from the permeability measurements with respect to
effective stress and shows an increasing trend with the quantity of gas adsorbed. This may
be related to the reduction in cleat porosity due to coal swelling. Coal's elastic modulus was
found to be larger at higher effective stress, which may be caused by its non-linear elastic
behaviour. The elastic modulus is also larger at lower temperatures, which may indicate
possible coal softening at higher temperatures. However, the relationship between elastic
modulus and gas adsorption is still unclear and will require further investigation.
2.5 Acknowledgement
The authors acknowledge the financial supports from the CSIRO Advanced Coal
Technology Portfolio, the Western Australia CSIRO-University Postgraduate Research
Scholarship, and National Research Flagship Energy Transformed Top-up Scholarship.
School of Mechanical and Chemical Engineering
Chapter 2 The University of Western Australia
2–17
2.6 References
Brace, W.F., Walsh, J.B. and Frangos, W.T., 1968. Permeability of granite under high pressure. Journal of
Geophysical Research, 73(6), 2225-2236.
Fitzgerald, J.E., Pan, Z., Sudibandriyo, M., Robinson, Jr. R.L., Gasem, K.A.M., and Reeves, S., 2005.
Adsorption of methane, nitrogen, carbon dioxide and their mixtures on wet Tiffany coal. Fuel. 84,
2351-2363.
Fujioka, M., Yamaguchi, S. and Nako, M., 2010. CO2-ECBM Field Tests in the Ishikari Coal Basin of Japan.
International Journal of Coal Geology, 82(3-4), 287-298.
Hall, F., Zhou, C., Gasem, K.A.M. and Robinson, Jr R.L., 1994. Adsorption of Pure Methane, Nitrogen, and
Carbon Dioxide and Their Mixtures on Wet Fruitland Coal. SPE 29194.
Mazumder, S., Karnik, A.A., and Wolf, K-H. A.A., 2006 Swelling of coal in response to CO2 sequestration
for ECBM and its effect on fracture permeability. SPE Journal, 11(3), 390-398.
Palmer, I., 2009. Permeability Changes in Coal: Analytical Modeling. International Journal of Coal Geology,
77(1-2), 119-126.
Pan, Z. and Connell, L.D., 2007. A theoretical Model for Gas Adsorption-induced Coal swelling.
International Journal of Coal Geology, 69, 243-252.
Pan, Z., Connell, L.D. and Camilleri, M., 2010. Laboratory Characterisation of Coal Reservoir Permeability
for Primary and Enhanced Coalbed Methane Recovery. International Journal of Coal Geology, 82(3-
4), 252-261.
Puri, R. and Yee, D., 1990. Enhanced Coalbed Methane Recovery, SPE Annual Technical Conference and
Exhibition. New Orleans, Louisiana.
Seidle, J.P., Jeansonne M.W. and Erickson D.J., 1992. Application of Matchstick Geometry To Stress
Dependent Permeability in Coals, SPE Rocky Mountain Regional Meeting. Casper, Wyoming.
Shi, J.Q. and Durucan, S., 2004. Drawdown Induced Changes in Permeability of Coalbeds: A New
Interpretation of the Reservoir Response to Primary Recovery. Transport in Porous Media, 56(1), 1-
16.
Sparks, D.P., McLendon, T.H., Saulsberry, J.L. and Lambert, S.W., 1995. The Effects of Stress on Coalbed
Reservoir Performance, Black Warrior Basin, U.S.A. SPE Annual Technical Conference and
Exhibition, Dallas, Texas.
Wong, S., Law, D., Deng, X., Robinson, J., Kadatz, B., Gunter, W.D., Ye, J., Feng, S., and Fan, Z., 2007.
Enhanced Coalbed Methane and CO2 Storage in Anthracitic Coals-Micro-pilot Test at South Qinshui,
Shanxi, China. International Journal of Greenhouse Gas Control, 1, 215-222.
van Bergen, F., Pagnier, H., and Krzystolik, P., 2006. Field Experiment of Enhanced Coalbed Methane– CO2
in the Upper Silesian Basin of Poland. Environmental Geosciences, 13, 201-224.
Zimmerman, R.W., Somerton, W.H., and King, S.M., 1986. Compressibility of porous rocks. Journal of
Geophysical Research, 91(B12), 12,765-12,777.
School of Mechanical and Chemical Engineering
Chapter 3 The University of Western Australia
3-1
Chapter 3
Linking Gas-Sorption Induced Changes in Coal Permeability to Directional Strains through a Modulus Reduction Ratio
Jishan Liua, Zhongwei Chena, Derek Elsworthc, Xiexing Miaob, Xianbiao Maob
a School of Mechanical and Chemical Engineering, The University of Western Australia, WA, 6009, Australia
b State Key Laboratory for Geomechanics and Underground Engineering, China University of Mining and
Technology, Jiangsu 221008, China
c Department of Energy and Mineral Engineering, Penn State University, PA 16802-5000, USA
Abstract: Although coal-gas interactions have been comprehensively investigated, prior
studies have focused on one or more component processes of effective stress or sorption-
induced deformation and for resulting isotropic changes in coal permeability. In this study
a permeability model is developed to define the evolution of gas sorption-induced
permeability anisotropy under the full spectrum of mechanical conditions spanning
prescribed in-situ stresses through constrained displacement. In the model, gas sorption-
induced coal directional permeabilities are linked into directional strains through an elastic
modulus reduction ratio, Rm. This defines the ratio of coal mass elastic modulus to coal
matrix modulus (0<Rm<1) and represents the partitioning of total strain for an equivalent
porous coal medium between the fracture system and the matrix. Where bulk coal
permeability is dominated by the cleat system, the portioned fracture strains may be used to
define the evolution of the fracture permeability, and hence the response of the bulk
aggregate. The coal modulus reduction ratio provides a straightforward index to link
anisotropy in deformability characteristics to the evolution of directional permeabilities.
Constitutive models incorporating this concept are implemented in a finite element model
to represent the complex interactions of effective stress and sorption under in-situ
conditions. The validity of the model is evaluated against benchmark cases for uniaxial
swelling and for constant volume reservoirs then applied to match changes in permeability
observed in a field production test within a coalbed reservoir.
Keywords: coal permeability anisotropy; coal swelling; coal-gas interactions; numerical
modelling; modulus reduction ratio
School of Mechanical and Chemical Engineering
Chapter 3 The University of Western Australia
3-2
3.1 Introduction
Knowledge of changes in coal permeability due to gas sorption-induced effective stress is
crucially important for the evaluation of both primary gas production from coalbed
reservoirs and for CO2-enhanced coalbed methane recovery (ECBM). For primary gas
production, the reduction of gas pressure increases effective stress which in turn closes the
fracture aperture and reduces the permeability. As the gas pressure reduces below the
desorption point, methane is released from the coal matrix to the fracture network and the
coal matrix shrinks. As a direct consequence of this matrix shrinkage the fractures dilate
and fracture permeability correspondingly increases. Thus, a rapid initial reduction of
fracture permeability (due to change in effective stress) is supplanted by a slow increase in
permeability (with matrix shrinkage). Whether the ultimate, long-term, permeability is
greater or less than the initial permeability depends on the net influence of these dual
competing mechanisms.
ECBM involves the injection of CO2 into a coal seam to promote the desorption of
coalbed methane (CBM) while simultaneously sequestering CO2 in the coal seam. This
process exploits the greater affinity of carbon dioxide (CO2) to adsorb onto coal relative to
methane (CH4), resulting in the net desorption of methane and its potential recovery as a
low-carbon fuel. Laboratory isotherm measurements for pure gasses have demonstrated
that coal can adsorb approximately twice (or more) as much CO2 (in moles) by volume as
methane (White et al., 2005). Correspondingly, CO2 injection with concurrent production
of methane can cause differential swelling of the coalbed particularly in the near wellbore
area. This may play an important role in determining the resulting deformation of the coal
matrix, the related permeability change and its impact on both gas diffusion to the cleats
and gas transport along the cleat network. Thus, the influence of these distinct but
connected changes in deformation, due to both effective stresses and to gas-sorption-
induced swelling, are key to unravelling the transient response to gas injection and recovery.
The complexity of the response is further increased by the overprinted effects of bedding
plane and cleat orientations, which together with directional stresses or displacement
restraints impart a further directional heterogeneity to the transient evolution of
permeability. Thus understanding the transient and anisotropic characteristics of
permeability evolution in fractured coals is of fundamental importance to the recovery of
methane from CBM reservoirs and equally important for CO2 storage suing ECBM.
School of Mechanical and Chemical Engineering
Chapter 3 The University of Western Australia
3-3
3.1.1 Experimental Observations
The potential impacts of coal swelling on the performance of coalbed methane production
and in the deep geological sequestration of CO2 have been investigated through
experimental, field-scale, and numerical studies.
Experiments have investigated the sorption and related swelling characteristics of coals.
The effects of water content on swelling and sorption have been explored for CO2 uptake
at 298 K (Ceglarska-Stefanska and Czaplinski, 1993) using a gas-flame coal, a gas-coking
coal and an anthracite and indicate a reduction in swelling strain for ―dry‖ coal versus ―pre-
wetted‖ samples (Ceglarska-Stefanska and Brzoska, 1998). Rates of swelling are controlled
largely by diffusive length scales imparted by the cleats. A surrogate of this case is
powdered coals where for powdered high volatile bituminous Pennsylvanian coals the
adsorption rate decreases with increasing grain size for all experimental conditions (Busch
et al., 2004). Similarly, coal type and rank (Robertson and Christiansen, 2005; Prusty, 2007)
influences the preferential sorption behavior and the evolution of permeability with these
changes is linked to macromolecular structure (Mazumder and Wolf, 2008). Adsorption
kinetics may also be determined for various gasses (e.g. for CO2 and CH4) using confining
cells to apply desired pressures and temperatures (Charrière et al., 2010) and using X-ray
CT methods to determine the resulting sorption isotherms (Jikich et al., 2009). These
experiments have focused on the isotropic characteristics of intact or powdered coals.
Conversely, some experiments have focused on the anisotropic characteristic of coal. Water
transmission characteristics have been shown to be significantly different (Gash et al., 1993)
under confining pressures when measured perpendicular to either face cleats, butt cleats, or
bedding planes. Directional flow experiments on isotropically compressed samples have
similarly confirmed the anisotropy of permeability for gas flows (Li et al, 2004). These
results are congruent with optical measurements of coal swelling under in CO2 and other
gasses where swelling in the plane perpendicular to the bedding plane was always
substantially higher than parallel to the bedding plane (Day, et al., 2008). This
phenomenon has also been observed in the field well tests in the Warrior Basin (USA)
where the anisotropy ratio of permeability in the direction of the bedding plane was as high
as 17:1 (Koenig and Stubbs, 1986).
School of Mechanical and Chemical Engineering
Chapter 3 The University of Western Australia
3-4
3.1.2 Permeability Models
Based on experimental observations, a variety of models have been formulated to quantify
the evolution of permeability during coal swelling/shrinking. The first attempts to quantify
the role of stresses on the evolution of coal-reservoir permeability assumed invariant
vertical stresses and linked changes in horizontal stress with the gas pressure and the
sorption strain (Gray, 1987). Permeability was computed as a function of reservoir
pressure-induced coal-matrix shrinkage assumed directly proportional to changes in the
equivalent sorption pressure. Since then, a number of theoretical and empirical
permeability models have been proposed. The Seidle-Huitt Model describes the evolution
of permeability assuming that all changes in permeability are caused by the sorption-
induced strain alone, neglecting the elastic strain (Seidle and Huitt, 1995). Another three of
the most widely used permeability models are the Palmer and Mansoori model (P-M
Model), the Shi and Durucan (S-D) model and the Advanced Resources International (ARI)
model (Palmer and Mansoori, 1996; Pekot and Reeves, 2002; Shi and Durucan, 2005). The
P-M model is strain-based so porosity change is defined by the change in volume strain,
and the related permeability change is calculated directly from that change in porosity. The
P-M relation is derived from the assumption of isotropic linear elasticity to determine
resulting changes in strain for an assumed invariant overburden stress. Consequently,
predicted changes in porosity are small and are permuted into permeability changes via a
cubic relationship between permeability and porosity. The S-D model is based on an
idealized bundled-matchstick geometry to represent a coalbed, and uses a stress-based
formulation to correlate changes in the effective horizontal stress caused by the volumetric
deformation together with the cleat or pore compressibilities. This stress-based model
accommodates changes in porosity and permeability through the anticipated change in
horizontal stress and includes no direct influence of swelling/shrinkage-induced strain.
Additionally, the Biot's coefficient is set to unity – requiring that the change in net stress is
equal to the difference between the overburden pressure and the change in pore pressure.
The ARI model describes the evolution of coal permeability using a semi-empirical
correlation to account for the changes of coal porosity due to pore compressibility and
matrix swelling/shrinkage (Pekot and Reeves, 2002). The ARI model is essentially
equivalent to the P-M model for saturated coal and where the strain versus stress
relationship fits the Langmuir isotherm (Palmer, 2009). However, although permeability
models incorporating sorption-induced effects have been widely studied, most of these
School of Mechanical and Chemical Engineering
Chapter 3 The University of Western Australia
3-5
studies are under the assumption of either an invariant total stress or uniaxial strain
conditions. These critical and limiting assumptions have been relaxed in new models
rigorously incorporating in-situ stress conditions (Zhang et al., 2008; Palmer, 2009; Connell,
2009) and are extended to rigorously incorporate CO2-CH4 coal-gas interaction relevant to
CO2-ECBM (Connell and Detounay, 2009; Chen et al., 2010).
Despite the complexity of models applied to represent the evolution of coalbed methane
reservoirs, few accommodate feedbacks of both anisotropy and coal-gas interactions on the
evolution of permeability – including the important roles of linked stress-deformation and
gas flow and adsorption/desorption processes. The effect of stress on the evolution of
flow anisotropy in orthogonally fractured media (Sayers, 1999) and in deformable granular
media (Du et al., 2004) has been investigated although not with the influence of gas
adsorption or desorption effects. The impact of permeability anisotropy and pressure
interference on CBM gas production has been investigated specifically to seek any unique
performance feature that might distinguish between isotropic or anisotropic permeability of
the CBM reservoir or to identify the drainage geometry (Chaianansutcharit, et al., 2001).
And analytical solutions have been presented for steady-state conditions with anisotropic
permeability (Al-Yousef, 2005). More recently, an alternative approach has been proposed
to develop an improved permeability model for CO2-ECBM recovery and CO2 geo-
sequestration in coal seams, integrating the textural and mechanical properties to describe
the anisotropy of gas permeability in coal reservoirs under confined stress conditions
(Wang et al., 2009).
3.1.3 This study
In this study, a novel permeability model is developed to define the evolution of gas
sorption-induced permeability anisotropy under in-situ stress conditions. Gas sorption-
induced coal directional permeabilities are linked to directional strains through the elastic
modulus reduction ratio (the ratio of coal mass elastic modulus to coal matrix modulus)
that represents the partition of the total strain for an equivalent porous coal medium
between the fracture system and the matrix. It is assumed that only the partitioned fracture
strains are responsible for the changes in directional permeabilities. These new relations
are the key cross couplings that link effective stress-related and sorption/desorption-related
changes in permeability to fluid pressure and gas content. These constitutive relationships
are incorporated into a finite element model to represent the complex interactions of stress
School of Mechanical and Chemical Engineering
Chapter 3 The University of Western Australia
3-6
and chemistry under in-situ conditions and to project their impact on rates and magnitudes
of gas recovery. The validity of the general model is evaluated against results for special
cases representing uniaxial swelling, constant volume reservoirs, and for the case of a
coalbed methane production well test. The incorporation of gas sorption-induced coal
permeability anisotropy into the multiphysics simulation of coal-gas interactions represents
a new and important contribution to this subject.
3.2 Approach
The overall approach is illustrated in Figure 3-1. The evaluation of fully coupled
deformation and gas transport in the fractured coal is conducted through four integrated
steps: (1) Coal deformation analysis; (2) Flow equivalence analysis; (3) Permeability
evolution analysis; and (4) Flow equivalence updating. These four steps are detailed in the
following sections.
Figure 3-1. Flow chart for evaluating coupled deformation and gas transport processes in
coal. Circled numbers represent steps of the analysis process.
1 Coal Deformation Analysis 2 Flow Equivalence Analysis
3 Permeability Evolution Analysis 4 Flow Equivalence Update
Coal Seam
Deformation Model
vzyx ,,,
Initial Condition
Isotropic coal permeability
0k
Discontinuous Model with
Uniform spacing s, and Aperture, b
Initial Condition
Anisotropic coal permeability
000 ,, zyx kkk
Discontinuous Model with
Uniform spacing s, and Nonuniform Aperture bx0,
by0, bz0
New Coal Permeability
zyx kkk ,,
Gas Flow & Transport
Model
sp ,
Gas Flow & Transport
Model
sp ,
1
2 2
3
4 4
2 2
School of Mechanical and Chemical Engineering
Chapter 3 The University of Western Australia
3-7
3.2.1 Coal Deformation Analysis
The mechanical properties of a discontinuous medium containing orthogonal fractures and
orthotropic response can be represented by the properties of an equivalent continuous
medium (Amadei and Goodman, 1981). The following assumptions are made:
(a) The coal is a homogeneous, isotropic and elastic continuum, and the system is
isothermal.
(b) Strains are infinitesimal.
(c) Gas contained within the pores is ideal, and its viscosity is constant under isothermal
conditions.
(d) Gas flow through the coal medium is assumed to be viscous flow obeying Darcy's
law.
According to the first assumption (a), the strain-displacement relation is expressed as
)(2
1,, ijjiij uu (3-1)
where ij is the component of the total strain tensor and iu is the component of the
displacement. The equilibrium equation is defined as
0, ijij f (3-2)
where ij denotes the component of the total stress tensor and if denotes the component
of the body force.
The gas sorption-induced strain s is assumed to result in only normal strains and these
resulting strains are isotropic. The effects of gas sorption on the deformation of coal seams
can be treated analogous to the effects of temperature for elastic porous media (e.g.,
Palmer and Mansoori, 1998), stress–strain relationships for an isothermal gas adsorbing
coalbed may be written as (Shi and Durucan, 2004)
ijs
ijijkkijij pKKGG
33
)9
1
6
1(
2
1 (3-3)
where mm ERE , )1(2
E
G , )21(3
E
K , mK
K1 ,
332211 kk.
School of Mechanical and Chemical Engineering
Chapter 3 The University of Western Australia
3-8
K represents the bulk modulus of coal and mK represents the bulk modulus of coal
matrixes. G is the shear modulus of coal, s is the sorption-induced strain, mE is the
Young's modulus of the matrix, mR is the modulus reduction ratio, E is the equivalent
Young's modulus of the coal-fracture assemblage and is the Poisson's ratio of the coal-
fracture assemblage. represents the Biot's coefficient, p the gas pressure in the pores
and ij is the Kronecker delta; 1 for ji and 0 for ji .
Combining Equations (3-1)–(3-3) yields the Navier-type equation expressed as
021
,,,
2
iisiii fKpeG
uG
(3-4)
where iu is the displacement in i direction, e is the volumetric strain, and
ip, is the partial
derivative of pore pressure with respect to i . Equation (3-4) is the governing equation
representing deformation of the continuum representation of the fractured coal allowing
deformations to be determined if fluid pressures, p, may be determined for both undrained
and drained response. Transient fluid pressures are recovered from the flow equation.
3.2.2 Flow and Transport Analysis
The mass balance equation for a single component gas is defined as
sgg Qt
m
)( q (3-5)
where g is the gas density, gq is the Darcy velocity vector and sQ is the gas source or
sink. m is the gas content including both free-phase and absorbed components (Saghafi, et
al., 2007) and is defined as
L
Lcgaffg
Pp
pVm
1 (3-6)
where ga is the gas density at standard conditions, c is the coal density and is
fracture porosity. LV represents the Langmuir volume constant and LP represents the
Langmuir pressure constant. According to the real gas law, the gas density is proportional
to the pore gas pressure and can be described as
School of Mechanical and Chemical Engineering
Chapter 3 The University of Western Australia
3-9
pZRT
M g
g (3-7)
where gM is the molar mass of gas, R is the universal gas constant, T is the gas
temperature and Z is the correction factor that accounts for the non-ideal behaviour of
the gas which changes with R and T.
Assuming that the effect of gravity is relatively small and can be neglected, the Darcy
velocity may be defined as
pk
g
q (3-8)
where denotes the dynamic viscosity of the gas and k denotes the permeability tensor,
expressed as
zzzyzx
yyyx
xzxyxx
kkk
kkk
kkk
k yz
(3-9)
Substituting Equations (3-6)–(3-9) into Equation (3-5), we obtain
s
f
L
Lac
L
LLacff Qpp
k
tPp
pVpp
t
p
Pp
PVp
2)(1 (3-10)
where ap is atmospheric pressure (1.0 atm or 101.325 kPa). In Equation (3-10), the
permeability k is dependent on the porosity, , while is a function of the volumetric
strain, and the sorption-induced strain, s , which will be illustrated in detail in the
following sections. Therefore, Equations (3-4) and (3-10) will be coupled through the
porosity-permeability relation and pore pressure evolution.
3.2.3 Coal Permeability Analysis
In the analysis of coal permeability the fractured coal mass is treated as a discontinuous
medium comprising both matrix and fractures (cleats), as illustrated in Figure 3-2. The
individual matrix blocks are represented by cubes and may behave isotropically with regard
to swelling/shrinking, thermal expansion, and mechanical deformability. The cleats are the
three orthogonal fracture sets and may also have different apertures and mechanical
properties ascribed to the different directions.
School of Mechanical and Chemical Engineering
Chapter 3 The University of Western Australia
3-10
Changes in coal permeability are determined by the redistribution of effective stresses or
strains due to changed conditions such as gas injection. Typically, stresses and strains will
evolve at different rates in the different Cartesian directions, i.e., x , y , z and
x , y , y , and result in anisotropic permeabilities, xk , yk , zk . To derive the
relationship between stresses/strains and directional permeability, several assumptions and
definitions are made:
The initial porous coal is substituted either by a discontinuous model with uniform fracture
spacing s and aperture b, as shown in Figure 3-2(a), or by a discontinuous model with
uniform spacing s and non-uniform aperture, 0xb , 0yb , and
0zb , as shown in Figure 3-
2(b).
Figure 3-2. Substitution of porous coal by discontinuous models.
The coal fracture porosity f can be determined by fracture spacing and aperture as
follows, sbsb
ssbf /3
)(
)(3
33
. For the two-dimensional case, the areal porosity for x-
or y-directions is defined as sbyfx /2 and sbxfy /2 , respectively.
Fracture and matrix deformation are both linear and fully recoverable, and deformations in
normal closure or opening are the predominant permeability alteration mode. Therefore,
coal permeability changes can be defined as a function of the variation of aperture in
corresponding directions i.e., xb , yb , and zb ; where the aperture variation partitioned
from the porous medium is realized through the elastic modulus reduction ratio, mR .
0xbs
x x
y
yz
z
0zb
0ybbs
x x
y
yz
z
(a) Initial Isotropic Permeability (b) Initial Anisotropic Permeability
School of Mechanical and Chemical Engineering
Chapter 3 The University of Western Australia
3-11
The coal matrix is functionally impermeable, and the dominant fluid flow within the
fractures may be defined by an equivalent parallel plate model. This enables the strain-
dependent effective porosity and the strain-dependent permeability field to be determined
if induced strains can be adequately partitioned between fracture and solid matrix.
The schematic diagram regarding the fracture aperture change and the effective stress
alteration is shown in Figure 3-3. The aperture closure induced by the effective stress
change can be calculated by
)(m
eieiii
Es
Esbb
(3-11)
Simplifying this equation, gives
Eb
EE
Esb ei
iei
m
i
)1( (3-12)
If assuming mm EER / , Eeiei / , then the above equation can be derived
eimi
fib
Rs
b
b
1
)1( (3-13)
Because sb , Equation (3-13) can be simplified into
eimi
fib
Rs
b
b
)1( (3-14)
wheremR is the elastic modulus reduction ratio,
ei is the effective strain change in the i-
direction, s is the fracture spacing and 0ib is the initial fracture aperture along i- direction.
Figure3-3. Schematic diagram of fracture aperture interaction with effective stress.
ei
Fractured Coal Equivalent Coal Coal Matrix
b
bs/2
s/2E mE
Fracture
Matrix
Matrix
u
mu
mi uub
ib
ei
ei
School of Mechanical and Chemical Engineering
Chapter 3 The University of Western Australia
3-12
Based on the above analysis, for the 3D case with three orthogonal sets of fractures, coal
directional permeability, xk , yk , and zk are defined as follows (Liu et al., 1999)
ji
ej
f
m
i
i R
k
k3
00
)1(31
2
1
(3-15)
ji
j
is
bk
12
3
0
0 (3-16)
where 0f is the initial fracture porosity at reference conditions, zyxji ,,, for 3D case
and yxji ,, for 2D case.
For the 2D case with two orthogonal sets of fractures, coal directional permeability, xk
and yk are defined as follows:
3
00
)1(21
ej
f
m
i
i R
k
k
, ji (3-17)
s
bk
j
i12
3
0
0 , ji (3-18)
Results from field and laboratory experiments indicate that coal permeability can change
significantly during adsorbable gas injection (e.g. CH4 and CO2). The injection gas pressure
tends to mechanically open coal cleats and thus enhance the permeability as the initial gas
pressure resides only in the fractures and any constrained change in total stress compresses
the matrix blocks. The subsequent gas adsorption into the coal matrix induces swelling
(volumetric strain) and has two effects: (1) it reduces effective stresses and causes an elastic
expansion of the coal due to injection gas pressure increase, overprinted by a (2) sorption-
induced swelling of the coal matrix as gas diffuses into coal matrixes. If expansion of the
cleat-matrix assemblage is constrained then fracture permeability reduces by narrowing and
even closing cleat apertures. When the coal swelling is taken into consideration, the total
effective strains in the Equation (3-15) can be replaced by the differences between the total
strain change, it , in i direction and the free swelling strain change, s , as follows
sitei 3
1 (3-19)
Thus for 3D case we have:
School of Mechanical and Chemical Engineering
Chapter 3 The University of Western Australia
3-13
ji
sjt
f
m
i
i R
k
k3
00 3
1)1(31
2
1
(3-20)
For the 2D case with two orthogonal sets of fractures, coal directional permeability, xk
and yk were defined as follows:
3
00 3
1)1(21
sjt
f
m
i
i R
k
k
, ji (3-21)
3.2.4 Coupled Model
For a system containing a single gas phase the sorption-induced volumetric strain s may
be represented by a Langmuir type function (Harpalani and Schraufnagel, 1990; Cui and
Bustin, 2005; Robertson and Christiansen, 2005), defined as
pP
p
L
Ls
(3-22)
where L and LP are the Langmuir-type matrix swelling/shrinkage constants, which
represent the maximum swelling capacity and the pore pressure at which the measured
volumetric strain is equal to L5.0 , respectively.
Substituting Equation (3-22) into Equation (3-4), we rewrite the governing equation for
deformation of the coal seam as
0)(21
,2,,
2
ii
L
LLiii fp
Pp
PKpe
GuG
(3-23)
From Equation (3-14), we can determine porosity f as
smff eRa
bb
)1(3
30
0
(3-24)
Then, the partial derivative of porosity f with respect to time can be expressed as
t
p
Pp
P
t
eR
t L
LLm
f
2)()1(3
(3-25)
Substituting Equation (3-25) into Equation (3-10) yields the governing equation for gas
flow through a coal seam with the effect of gas sorption incorporated as
School of Mechanical and Chemical Engineering
Chapter 3 The University of Western Australia
3-14
s
L
Lacm
L
LL
L
Lacm
L
LLacff
Qt
e
Pp
pVppR
ppk
t
p
Pp
P
Pp
pVppR
Pp
PVp
)1(3
-)(
)1(3)(
122
(3-26)
In our analysis of coal deformation, the fractured coal mass is represented as an
orthotropic fractured medium which is replaced as an equivalent continuous medium.
When we conduct flow analysis, we partition the total effective strain from the equivalent
medium between coal matrix and the fracture system. Only the partitioned strain for the
fracture system contributes to the permeability change. When the rock mass reduction ratio
is unity, i.e. 1mR then the equivalent modulus of the fractured medium is equal to that
of the coal matrix. In other words the coal mass may be considered as unfractured or the
fractures are infinitely small. Conversely, in the limit as 0mR then the coal matrix is
infinitely stiff and the observed deformational response is equivalent to that of the
fractures alone. Therefore the parameter mR1 represents the ratio of the partitioned strain
for the fracture system to the total equivalent strain. If 1mR , the partitioned strain for
the fracture system is due to that of the matrix modulus, which is essentially zero in
comparison to the anticipated response of the more compliant fractures; therefore a very
small permeability change due to the deformation of the matrix results and this is taken as
null. If 0mR then the partitioned strain is predominantly due to the fracture deformation.
Where the fractures are typically more compliant than the host from which they are derived,
then a maximum permeability change results.
The total effective strain is the difference between the total strain, as determined by the
constrained boundary conditions, and the free swelling strain. Therefore, the boundary
conditions also control the evolution of coal permeability.
In the following sections, we use three examples to illustrate these principles. These are
respectively conditions of (1) uniaxial strain; (2) constant reservoir volume; and (3) the
behaviour of a field case.
3.3 Uniaxial Strain Condition
For the case of uniaxial strain, the lateral strains, tx and ty , are equal to zero. Based
on Hooke's law, the relation between stress and strain increments are:
School of Mechanical and Chemical Engineering
Chapter 3 The University of Western Australia
3-15
sezexstxexE
3
11
1
3
1 (3-27)
sezeystyeyE
3
11
1
3
1 (3-28)
sexezez
E
pp
E
13
2
1
2112
1 0 (3-29)
Substituting these equations into Equation (3-20) gives
3
0
0
3
000
13
2
1
211)1(31
2
1
3
1)1(31
2
1
s
f
m
s
f
m
y
y
x
x
E
ppR
R
k
k
k
k
(3-30)
3
00 3
1)1(31
s
f
m
z
z R
k
k
(3-31)
As shown in Equations (3-30) and (3-31), coal permeability in the x-direction is not equal
to the permeability in the z-direction. The vertical permeability, zk , is determined by the
swelling strain only while the horizontal permeability, yx kk or , is determined both by the
swelling strain and by the mechanical deformation. It is obvious that zk is not equal to xk
even if 000 zyx kkk . In order to illustrate these conclusions graphically, we use a set of
parameters in the Table 3-1 to quantify the directional permeabilities. The calculated
results are shown in Figure 3-4. If 1mR , the partitioned strain for the fracture system is
zero; therefore, no permeability change is induced: permeability ratios are equal to unity; If
0mR , the partitioned strain for the fracture system is 100%; therefore, maximum
permeability changes zx kk and are induced.
Table 3-1. Parameters used for the example calculations
Parameter Definition value
E Young's modulus 2.7GPa
Poisson's ratio 0.4
b Fracture aperture 0.0001
s Fracture spacing 0.01
L Langmuir strain constant 0.03
Lp
Langmuir pressure constant 3MPa
School of Mechanical and Chemical Engineering
Chapter 3 The University of Western Australia
3-16
mR
Elastic modulus reduction ratio 0, 0.5, 1
Figure 3-4. Coal permeability as a function of matrix pore pressure under uniaxial strain
condition and the influence of gas adsorption. Permeability ratios are regulated by the
modulus ratio, mR : 1mR represents no fracture influence; 0mR represents fracture
influence only; 5.0mR combined influence of fracture and matrix deformation.
3.4 Displacement Controlled Condition
For the displacement controlled (or constant reservoir volume) case, strains in all directions,
tx , ty , and tz are equal to zero. Substituting zero value into Equation (3-20) gives
3
0000 3
1)1(31
s
f
m
z
z
y
y
x
x R
k
k
k
k
k
k
(3-32)
As shown in Equation (3-32), coal permeability ratio in the x- and y-directions is equal to
the permeability ratio in the z-direction. All permeability ratios are determined by the
swelling strain only. It is apparent that zk is equal to xk if 000 zyx kkk . In order to
illustrate these conclusions graphically, we use the parameters in the Table 3-1 to quantify
the directional permeabilities. The calculated results are shown in Figure 3-5. We also
0
0
0
1
5.0
0
x
xm
x
xm
x
xm
k
kR
k
kR
k
kR
0
0
0
1
5.0
0
z
zm
z
zm
z
zm
k
kR
k
kR
k
kR
Matrix Pore Pressure, MPa
Pe
rme
abili
ty R
atio
0.01
0.1
1
10
0 1 2 3 4 5 6 7
School of Mechanical and Chemical Engineering
Chapter 3 The University of Western Australia
3-17
compare the results for the constant volume case with the ones for the uniaxial strain case
in z-direction. For the constant volume case, 100% of the swelling strain contributes to the
total effective strain; for the uniaxial strain case, only a portion of the swelling strain
contributes to the total effective strain due to the unconstrained condition in the vertical
direction. Therefore, maximum permeability changes are induced under the displacement
controlled case.
Figure 3-5. Coal permeability as a function of matrix pore pressure under constant
reservoir volume and uniaxial strain conditions. Permeability ratios are equal in all
directions and regulated by the modulus ratio, mR : 1mR represents no fracture
influence; 0mR represents fracture influence only; 5.0mR combined influence of
fracture and matrix deformation. A cut-off permeability limit is applied for 0mR .
3.5 Field Case
It is generally believed that the in-situ response of a coal gas reservoir to gas production
(injection) can be approximated either by the uniaxial deformation-permeability model (Liu
and Rutqvist 2010) or by the constant volume-permeability model (Massarotto, 2009). In
this section, we apply both coal permeability models to a field case.
Mavor and Vaughn (1997) reported coal permeability results of three wells in the Valencia
Canyon area of the San Juan Basin, and found coal permeability increased between 2.7 and
0.0001
0.001
0.01
0.1
1
0 1 2 3 4 5 6 7
1mR
Strain Uniaxial5.0mR
Strain Uniaxial0mR
VolumeConstant 5.0mR
VolumeConstant 0mR
Matrix Pore Pressure, MPa
Pe
rme
abili
ty R
atio
School of Mechanical and Chemical Engineering
Chapter 3 The University of Western Australia
3-18
7.1 times as gas pressure decreased. The initial gas pressures for these wells are 5.35, 6.60
and 6.41MPa, respectively. Because of the small differences, an average value of 6.12MPa
is used in this evaluation. The following mechanical properties and matrix swelling
parameters are taken directly from Liu and Rutqvist (2010): 3.0 , 2900E MPa,
55.2Lp MPa, 0043.0L . The initial fracture porosity is 0.05% (Mavor and Vaughn,
1997). These values are representative of the San Juan Basin coalbed.
For the uniaxial strain assumption, Equations (3-30) and (3-31) were used to evaluate the
permeability changes. In this case, the vertical permeability is different from the horizontal
permeability. This means that coal shrinkage induces the permeability anisotropy. In our
analysis, we match the horizontal permeabilities with the field data. For the constant
reservoir volume assumption, Equation (3-32) was used to evaluate the permeability
changes.
For both assumed conditions, only the modulus reduction ratio,
mR , is adjustable. Best
matches were achieved when 4.0mR and 6.0mR , respectively, as shown in Figure 3-6.
For the uniaxial deformation case, 6.01 mR , representing a partitioned strain of the
equivalent porous coal medium for the fracture system of 60%. For the constant volume
case, 4.01 mR , representing a partitioned strain of the equivalent porous coal medium
for the fracture system of 40%.
1
10
100
0 1 2 3 4 5 6 7
0k
k x
00 k
k
k
k xz
000 k
k
k
k
k
k zx
Case VolumeConstant 4.01
CaseStrain Uniaxial6.01
%05.03.02900
12.60043.055.2
0
0
m
m
LL
R
R
MPaE
MPapMPap
Pore Pressure, MPa
Pe
rme
abili
ty R
atio
p
p
School of Mechanical and Chemical Engineering
Chapter 3 The University of Western Australia
3-19
Figure 3-6. Matching with field data through use of the uniaxial strain and constant
reservoir volume models.
3.6 Evaluation of Coupled Processes
A field scale model is used to simulate the performance of coalbed methane production
under in-situ conditions. Input parameters for this simulation are identical to the
parameters used in Section 3.5. The simulation model geometry is 200m by 200m with a
methane production well located at the lower left corner. For the coal deformation model,
all four sides are constrained in the normal direction. For the gas transport model, the coal
is saturated initially with CH4 and the initial pressure is 6.12MPa. A condition of
atmospheric pressure is applied at the boundary representing the production well.
Simulation results are presented in Figures 3-7 through 3-9.
Figure 3-7. Spatial and temporal evolution of coal permeability ratios on a diagonal radial
traverse from the production well.
1
10
100
0 20 40 60 80 100
Pe
rme
ab
ility
ra
tio
, k/k
0
Distance from Production well (m)
dayst 100
dayst 1000
dayst 5000
dayst 10000
6.0
0043.0
55.2
12.6
3.0
2900
005.0
8.0
0
0
m
L
L
f
R
MPap
MPap
MPaE
bp
100m
10
0m
School of Mechanical and Chemical Engineering
Chapter 3 The University of Western Australia
3-20
Figure 3-8. Evolution of coal permeability at a specific evaluation point.
Figure 3-9. Evolution of the cumulative gas production and pore pressure.
In this simulation, the reservoir volume remains unchanged throughout the production.
This assumption requires that global strains within the coal seam are zero. However, this
constraint does not preclude desorption-induced shrinkage (or swelling) of individual coal
blocks and complementary opening (or closing) of fractures. The direct consequence of
0
5
10
15
20
25
30
35
40
0.1 1 10 100 1000 10000
Pe
rme
ab
lity ra
tio
, k/k
0
Time (day)
6.0
0043.0
55.2
12.6
3.0
2900
005.0
8.0
0
0
m
L
L
f
R
MPap
MPap
MPaE
bp
100m
10
0m
Evaluation
Point
0
1
2
3
4
5
6
7
0
1
2
3
4
5
6
7
0 500 1000 1500 2000 2500 3000
Po
re p
ressu
re (M
Pa
)
Ga
s p
rod
uctio
n (×
10
5 m
3)
Time (day)
6.0
0043.0
55.2
12.6
3.0
2900
005.0
8.0
0
0
m
L
L
f
R
MPap
MPap
MPaE
bp
100m
10
0m
School of Mechanical and Chemical Engineering
Chapter 3 The University of Western Australia
3-21
these internal transformations is the isotropic change in permeability defined through
Equation (3-32). As shown in this equation, the change in coal permeability is defined only
by the swelling strain. This represents the ideal case, i.e., 100% of the swelling strain
contributes to the effective stress-induced coal deformation. However, only a portion of
the effective stress-induced coal deformation contributes to the permeability change as this
is modulated through the parametermR1 . In this case, mR1 is equal to 0.4. This
means that only ~40% of the total effective stress-induced coal strain (equal to the swelling
strain) is directly responsible for the permeability growth, as shown in Figures 3-7 and 3-8.
Figure 3-9 shows the relation between the cumulative gas production and the pore pressure.
3.7 Conclusions
A novel permeability model has been developed to define the evolution of gas sorption-
induced permeability anisotropy under in-situ stress conditions. This was implemented into
a fully coupled finite element model of coal deformation and gas transport in a coal seam.
Based on the model evaluations and the analysis of coupled processes, the model
adequately and consistently reflects the conceptual assumptions:
The directional permeability of coal is determined by the mechanical boundary
conditions, the ratio of coal bulk modulus to coal matrix modulus, the initial
fracture porosity, and the magnitude of the gas-pressure-induced coal swelling
strain. The boundary conditions control the magnitudes of total strains while the
modulus reduction ratio partitions the effective strain (total strain minus the
swelling strain) between fracture and matrix.
For restraint conditions of uniaxial strain and for a constant volume reservoir,
changes in coal permeability are determined only by the gas pore pressure and the
swelling strain. In both cases, the influence of effective stress is absent in the
permeability models.
Analysis including the effect of the fully coupled processes illustrates how coal
permeability evolves both in space and in time. These evolutions are the direct
outcomes of feedbacks of coal-gas interactions on the evolution of permeability,
stress deformation, gas flow and adsorption/desorption processes.
School of Mechanical and Chemical Engineering
Chapter 3 The University of Western Australia
3-22
3.8 Acknowledgements
This work was supported by the Western Australia CSIRO-University Postgraduate
Research Scholarship, National Research Flagship Energy Transformed Top-up
Scholarship, and by NIOSH under contract 200-2008-25702. These various sources of
support are gratefully acknowledged.
3.9 References
Al-Yousef, H.Y., 2005. Permeability anisotropy measurement on whole cores—Analytical solution and
application. SPE 93559.
Amadei, B., Goodman, R. E., 1981. A 3D constitutive relation for fractured rock masses, Proceedings of the
International Symposium on the Mechanical Behavior of Structured Media, Ottawa, ON, pp. 249-268.
Busch, A., Gensterblum, Y., Krooss, B.M., Littke, R., 2004. Methane and carbon dioxide adsorption–
diffusion experiments on coal: upscaling and modeling. International Journal of Coal Geology 60,
151-168.
Ceglarska-Stefanska, G. and Brzoska, K., 1998. The effect of coal metamorphism on methane desorption.
Fuel 77(6), 645-648.
Ceglarska-Stefanska, G. and Czaplinski, A., 1993. Correlation between sorption and dilatometric processes in
hard coals. Fuel 72(3), 413-417.
Chaianansutcharit, T., Chen, H.Y., Teufel, W.L., 2001. Impacts of permeability anisotropy and pressure
interference on coalbed methane (CBM) production. SPE 71069.
Charrière, D., Pokryszka, Z., Behra, P., 2010. Effect of pressure and temperature on diffusion of CO2 and
CH4 into coal from the Lorraine basin (France). International Journal of Coal Geology 81(4), 373-380.
Chen, Z.W., Liu, J.S., Elsworth, D., Connell, D.L., Pan Z.J., 2010. Impact of CO2 injection and differential
deformation on CO2 injectivity under in-situ stress conditions. International Journal of Coal Geology
81 (2), 97-108.
Cui, X., and Bustin, R. M., 2005. Volumetric strain associated with methane desorption and its impact on
coalbed gas production from deep coal seams, The American Association of Petroleum Geologists:
Bulletin 89 (9)1181-1202.
Day, S., Fry, R., Sakurovs, R., 2008. Swelling of Australian coals in supercritical CO2. International Journal of
Coal Geology 74(1), 41-52.
Du, J.C., Wong, C.K.R., Choy, E., 2004. Effects of strain-induced anisotropy in permeability on the
deformation-flow-heat transfer in porous media. The 6th North America Rock Mechanics
Symposium, Houston, Texas. ARMA/NARMS 04-613.
Harpalani, S., and Schraufnagel, A., 1990. Measurement of parameters impacting methane recovery from coal
seams, International Journal of Mining and Geological Engineering 8, 369-384.
Gash B.W., Volz, R.F., Potter, G., and Corgan, J.M., 1993. The effects of cleat orientation and confining
pressure on cleat porosity, permeability and relative permeability in coal. The 1993 International
Coalbed Methane Symposium. The University of Alabama, Tuscaloosa, Alabama, USA, paper 9321.
Gray, I., 1987. Reservoir engineering in coal seams: Part 1- the physical process of gas storage and movement
in coal seams. SPE Reservoir Engineering 2(1), 28-34. SPE-12514-PA.
Jikich, A. S., McLendon, R., Seshadri, K., Irdi, G., Smith, H.D., 2009. Carbon dioxide transport and sorption
behavior in confined coal cores for carbon sequestration. SPE Reservoir Evaluation & Engineering
12(1), 124-136.
School of Mechanical and Chemical Engineering
Chapter 3 The University of Western Australia
3-23
Koenig, R.A. and Stubbs, P.B., 1986. Interference testing of a coalbed methane reservoir. The SPE
Unconventional Gas Technology Symposium. The Society of Petroleum Engineers, Richardson,
Texas, USA. SPE 15225.
Li, H.Y., Shimada, S., Zhang, M., 2004. Anisotropy of gas permeability associated with cleat pattern in a coal
seam of the Kushiro coalfield in Japan. Environmental Geology 47, 45-50.
Liu, H.-H., and Rutqvist, J., 2010. A new coal-permeability model: Internal swelling stress and fracture–matrix
interaction. Transport in Porous Media. 82, 157–171.
Liu, J.S., Elsworth, D., Brady, B.H., 1999. Linking stress-dependent effective porosity and hydraulic
conductivity fields to RMR. International Journal of Rock Mechanics and Mining Sciences 36, 581-
596.
Massarotto, P., Golding, D.S., Rudolph, V., 2009. Constant volume CBM reservoirs: An important principle.
2009 International Coalbed & Shale Gas Symposium. Paper No.0926.
Mavor, M. J. and Vaughn, J.E., 1997. Increasing absolute permeability in the San Juan basin Fruitland
formation, Proc. 1997 Int. Coalbed Methane Symp., Tuscaloosa, Alabama, 12–16 May, 33–45.
Mazumder, S., Wolf, K.-H., 2008. Differential swelling and permeability change of coal in response to CO2
injection for ECBM. International Journal of Coal Geology 74 (2), 123-138.
Palmer, I., 2009. Permeability changes in coal: analytical modeling. International Journal of Coal Geology 77,
119-126.
Palmer, I., Mansoori, J., 1996. How permeability depends on stress and pore pressure in coalbeds: a new
model. SPE-52607.
Pekot, L.J., Reeves, S.R., 2002. Modeling the effects of matrix shrinkage and differential swelling on coalbed
methane recovery and carbon sequestration. U.S. Department of Energy DE-FC26-00NT40924.
Promeroy, C. D. and Robinson, D. J., 1967. The effect of applied stresses on the permeability of a middle
rank coal to water. International journal of rock mechanics and mining sciences 4, 329-343.
Prusty, B.K., 2007. Sorption of methane and CO2 for enhanced coalbed methane recovery and carbon
dioxide sequestration. Journal of Natural Gas Chemistry 17, 29-38.
Robertson, E.P., Christiansen, R.L., 2005. Measurement of sorption-induced strain. International Coalbed
Methane Symposium, University of Alabama, Tuscaloosa. Paper 0532.
Robertson, E.P., Christiansen, R.L., 2007. Modeling laboratory permeability in coal using sorption-induced-
strain data. SPE Reservoir Evaluation & Engineering 10(3), 260–269.
Saghafi, A., Faiz, M. and Roberts, D., 2007. CO2 storage and gas diffusivity properties of coals from Sydney
Basin, Australia. International Journal of Coal Geology 70, 240-254.
Sayers, C. M., 1990. Stress-induced fluid flow anisotropy in fractured rock. Transport in Porous Media 5, 287-
297.
Seidle, J.P., Huitt, L.G., 1995. Experimental measurement of coal matrix shrinkage due to gas desorption and
implications for cleat permeability increases. SPE-30010-MS.
Shi, J.-Q., Durucan, S., 2005. A model for changes in coalbed permeability during primary and enhanced
methane recovery. SPE Reservoir Evaluation & Engineering 8(4), 291–299.
Wang, G.X., Massarottoa, P., Rudolpha, V., 2009. An improved permeability model of coal for coalbed
methane recovery and CO2 geosequestration. International Journal of Coal Geology 77, 127-136.
White, C.M., Smith, D.H., Jones, K.L., et al., 2005. Sequestration of carbon dioxide in coal with enhanced
coalbed methane recovery: a review. Energy Fuels 19(3), 659–724.
Zhang, H.B., Liu, J., Elsworth, D., 2008. How sorption-induced matrix deformation affects gas flow in coal
seams: a new FE model. International Journal of Rock Mechanics and Mining Sciences 45(8), 1226-
1236.
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-1
Chapter 4
Effect of the Effective Stress Coefficient and Sorption-Induced Strain on the Evolution of Coal Permeability: Experimental Observations
Zhongwei Chena,b, Zhejun Panb, Jishan Liua, Luke D. Connellb, Derek Elsworthc
a School of Mechanical and Chemical Engineering, The University of Western Australia, WA 6009, Australia
b CSIRO Earth Science and Resource Engineering, Private Bag 10, Clayton South, Victoria 3169, Australia
c Department of Energy and Mineral Engineering, Penn State University, PA 16802-5000, USA
Abstract: Permeability is one of the most important parameters for CO2 injection in coal
to enhance coalbed methane recovery. Laboratory characterization of coal permeability
provides useful information of in situ permeability behaviour of coal seams when
adsorbing gasses such as CO2 are injected. In this study, a series of experiments have been
conducted for coal samples using both non-adsorbing and adsorbing gasses at various
confining stresses and pore pressures. Our observations have showed that even under
controlled stress conditions, coal permeability decreases with respect to pore pressure
during the injection of adsorbing gasses. In order to find out the causes of permeability
decrease for adsorbing gasses, a non-adsorbing gas (helium) is used to determine the
effective stress coefficient. In these experiments using helium, the impact of gas sorption
can be neglected and any permeability reduction is considered as due to the variation in the
effective stress, which is controlled by the effective stress coefficient. The results show that
the effective stress coefficient is pore pressure dependent and less than unity for the coal
samples studied. The permeability reduction from helium experiments is then used to
calibrate the subsequent flow-through experiments using adsorbing gasses, CH4 and CO2.
Through this calibration, the sole effect of sorption-induced strain on permeability change
is obtained for these adsorbing gas flow-through experiments. In this paper, experimental
results and analyses are reported including how the impact of effective stress coefficient is
separated from that of the sorption-induced strain on the evolution of coal permeability.
Keywords: effective stress coefficient; swelling and shrinking; adsorption; CO2 storage
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-2
4.1 INTRODUCTION
Knowledge of changes in coal permeability due to coal matrix swelling/shrinkage strain is
crucial for the evaluation of both primary gas production from coal reservoirs and for CO2-
enhanced coalbed methane recovery (ECBM) (van Bergen et al., 2009a,b). For primary gas
production, as the gas pressure decreases below the desorption pressure, methane is
released from the coal matrix to the fracture network and the coal matrix shrinks. As a
direct consequence of this matrix shrinkage, the fractures dilate and fracture permeability
correspondingly increases. Thus, a rapid initial reduction of fracture permeability (due to
decrease of pore pressure) is supplanted by a slow increase in permeability at later
production stage (due to matrix shrinkage). Whether the ultimate, long-term permeability is
greater or less than the initial permeability depends on the net results of these two
competing mechanisms. ECBM involves the injection of CO2 into coal seams to displace
methane recovered as an energy source, while providing the additional benefit of reducing
greenhouse gas emissions by storing the CO2 underground (White et al., 2005; Liu et al.,
2010b).
4.1.1 Experiments on coal swelling/shrinkage and permeability change
Coal swelling/shrinkage due to gas adsorption/desorption is a well-known phenomenon
and is regarded as a key component for coal reservoir permeability behaviour during
primary and enhanced coalbed methane recovery (e.g. Palmer and Mansoori, 2009; Shi and
Durucan, 2004). Measurements of the effects of gas desorption on coal volumetric strain
have been performed for the injection of different gasses. The implications for changes in
cleat permeability have been evaluated using a matchstick geometry model (Harpalani and
Schraufnagel, 1990; Palmer and Mansoori, 1996; Seidle et al., 1992; Seidle and Huitt, 1995;
Shi and Durucan, 2004; St. George and Barakat, 2001). Laboratory measurements of coal
swelling with gas sorption and the causes of swelling have been investigated by various
researchers (Bustin et al., 2008; Chikatamarla et al., 2004; Cui et al., 2007; Day et al., 2008;
Levine, 1996; Moffat and Weale, 1955; Pan and Connell, 2007; Pan and Connell, 2011;
Reucroft and Patel, 1986; Reucroft and Sethuraman, 1987; Robertson and Christiansen,
2005; St. George and Barakat, 2001; Wang et al., 2010). These studies investigated the
change in volumetric strain rate as a result of gas pressure change, and suggest that
swelling/shrinkage induced deformations dominate over the effective-stress-generated
deformations at low gas pressures for both carbon dioxide and methane.
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-3
Many experiments have been completed to evaluate the separate influences of the effective
stresses and sorption-induced strains on the evolution in permeability (Harpalani and Chen,
1997; Karacan, 2003; Mazumder et al., 2006; Pan et al., 2010a; Pini et al., 2009; Robertson
and Christiansen, 2007; Siriwardane et al., 2009). Patching (1965) found that the
permeability measured using carbon dioxide was somewhat less than the permeability by air
or nitrogen. Somerton et al. (1975) investigated the permeability of fractured coal using
methane and also observed that permeability measured using methane was 20–40% lower
than that using nitrogen. Investigations of the impact of adsorption and effective stress on
permeability change have shown an increase in permeability with decreasing effective stress
on the sample using non-adsorbing gas and a reduction in permeability caused by swelling
using an adsorbing gas (Pini et al., 2009; Pan et al., 2010a). Nevertheless, an increase of
permeability even with the injection of adsorbing gasses has also been observed
(Mazumder et al., 2006; Robertson and Christiansen, 2007; Mazumder and Wolf, 2008). In
addition, Karacan (2003) used X-ray CT technology to reveal the kinetics of the
heterogeneous processes occurring in a consolidated coal maintained under constant
pressure difference as the CO2 injection pressure and confining pressure were increased to
different levels. Lin et al. (2008) conducted a set of experiments under constant effective
stress condition (the difference between confining stress and pore pressure was kept
unchanged throughout the tests) with the injections of CO2, N2, and their binary mixture.
Gas composition and pressure influence on coal porosity and permeability was also
conducted by Mavor and Gunter (2006). The permeability decreases with escalating pore
pressure has been observed. The rebound of coal permeability with decreasing the pore
pressure under constant net effective stress was observed as well (Huy et al., 2010; Liu et al.,
2010a; Pini et al., 2009; Siriwardane et al., 2009). With the assumption of effective stress
coefficient to be unity, many experiments of moisture effect on coal physical properties
and gas sorption rate and sorption capacities have also been conducted, and a general
reduction trend of these parameters with increasing moisture has been observed (Clarkson
and Bustin, 2000; Gash, 1991; Pan et al., 2010b; Ozdemir and Schroeder, 2009).
4.1.2 Mesurements of effective stress coefficient
The law of effective stress was first presented by Terzaghi (1923) to explain the
consolidation of saturated soil and the interaction between fluid and soil. Including the
effective stress coefficient, the law of effective stress is (Biot, 1941).
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-4
ijijij p ' (4-1)
where is the effective stress coefficient which ranges from 0 to 1, expressed as
mKK1 ; K is the bulk skeletal modulus of the rock media and mK is the modulus of
rock matrix (grain modulus).
The effective stress coefficient, in a cross-section in the porous material, denotes the ratio
of the area occupied by the fluid to the total area (Bear, 1972). This coefficient describes to
which extent the fluid pressure counteracts elastic deformation of porous media (Alam et
al., 2010). For the granular soil (as shown in Figure 4-1), the contact area among grains is
very small, so any cross-section can be replaced by a curved face, so the corresponding
effective stress coefficient can be approximately assumed to be 1.0. However, for porous
rock (i.e. coal) composed of crystallization or cementation as shown in Figure 4-2, the
curved face, similar to that in Figure 4-1, does not exist. Thus the effective stress
coefficient is less than 1.0 (Zhang et al., 2009).
Figure4-1. Sketch of effective stress coefficient 1 for soil and soil-like materials (Zhang
et al., 2009).
Figure4-2. Sketch of pore structure in rock or coal media (Zhang et al., 2009).
Many researchers have studied the effective stress coefficient, but almost exclusively for
rocks. For instance, Walsh (1981) showed that α= 0.9 for a rock mass containing a polished
joint, and Kranzz (1979) defined α = 0.56 for rock containing a tensile joint. A linear
Cross section
Curved face
(a) Distribution of grains
(b) Cross section (c)Curved face
Pore
Solid
Arbitrary
section
(a) Pore distribution (b) Arbitrary section
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-5
increase of effective stress coefficient with pressure difference (defined as the difference
between confining stress and pore pressure) has also been observed (Ghabezloo et al.,
2009). Nevertheless, only a few experiments have been carried out on coal, even though
the importance of this parameter has been widely realized on both the primary gas
production of coal reservoirs and the predictions of gas outburst. Experimental
measurements on different coals have suggested that effective stress coefficient is not a
constant, and is a bilinear function of volumetric stress and pore pressure of coal
(Yangsheng et al., 2003), implying a change in bulk modulus or pore modulus with pore
pressure. Similarly, a series of loading-unloading cycles applied by gas pressure were
performed on coal samples using a non-adsorbing helium gas and a relationship was
established between the effective stress and gas pressure with 71.0 (St. George and
Barakat, 2001). Nevertheless, these determinations of the effective stress coefficient are
based on the measured bulk volumetric change, which means that volumetric strain is the
only physical property required for its calculation.
However, Robin (1973) considered the variation of permeability, k , instead of volumetric
strain, and derived another effective stress equation. In Robin's work, based on the pore
volume change, the following relation was used (Nur and Byerlee, 1971):
mKKK
11111
(4-2)
where K is the bulk modulus for pore system, and is porosity. K and
mK are porous
media bulk modulus and matrix bulk modulus respectively.
If porosity is very small (e.g. %10 ), the above equation can be simplified into:
)( KK
KKK
m
m
(4-3)
Similar to derivation of Biot's coefficient (i.e. Equation (4-1)), the effective stress
coefficient for the variation of permeability (or pore volume) can be expressed as:
ij
m
ije pKK
K
1 (4-4)
Therefore, the effective stress coefficient in this case can be defined as:
KKK m 1 (4-5)
It should be noted that both volumetric strain and the variation in permeability could be
used to quantify ‗elastic properties‘ of the media. As mentioned above, α is calculated on
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-6
the observed variation in bulk volumetric strain change. However, in this work, the
measurements are based on the observed variation in permeability and the measured value,
β, is also called the effective stress coefficient.
Based on the variation in permeability, Nur and Byerlee (1971) obtained 64.0 and
97.0 for the Weber sandstone. Bernabe (1987) studied the permeability of several
crystalline rocks and found that effective stress coefficient decreases with increasing
confining pressure due to the changes in the geometry of the cracks during closure.
Measurement of effective stress coefficient for coal is rare. Nevertheless, it is important
since coal is a weak rock and the grain compressibility is larger than expected and may have
a significant impact (Palmer, 2009; Zheng, 1993).
4.1.3 Objective of this study
In this work, a series of experiments are conducted using both non-adsorbing and
adsorbing gasses all under the constant pressure difference condition, which is defined as
the differential pressure between confining stress and pore pressure. Firstly, the effective
stress coefficient is obtained from flow-though experiments with a non-adsorbing gas
(helium). In these experiments, the impact of sorption induced strain is negligible.
Furthermore, Klinkenberg effect can be neglected at high pressures. Thus, any permeability
reduction is considered to be due solely to the variation in the effective stress coefficient.
Secondly, the permeability reduction from the experiments conducted with the non-
adsorbing gas is used to calibrate the subsequent experiments using adsorbing gasses
(carbon dioxide and methane). As a result, the effect of sorption-induced strain on
permeability change is obtained.
4.2 Experimental methods
4.2.1 Experiment apparatus
The schematic of the Triaxial Multi-Gas Rig used for this work is shown in Figure 4-3. The
rig is developed to measure gas adsorption, swelling, permeability for coal cores. Gas is
injected from the upstream injection pump to the sample. After the coal sample reaches
adsorption equilibrium, which usually takes a few days to a few weeks, transient method is
applied to measure permeability. The upstream cylinder is charged to about 30kPa higher
than the sample pressure while the downstream cylinder pressure is about 30kPa lower
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-7
than the sample pressure. Then gas is allowed to flow through the sample from the
upstream cylinder to the downstream cylinder. Permeability can be calculated from the
pressure decay curve measured by a differential pressure transducer installed between the
upstream and downstream cylinders. The calculation of permeability is described in the
next section. Radial and axial displacements are measured at each adsorption step to obtain
swelling/shrinkage strain. The sample cell and other parts of the rig are in a temperature
controlled cabinet to maintain constant temperature during the experiment. The core
sample is wrapped with a thin lead foil then a rubber sleeve before it is installed in the cell.
The thin lead foil is to prevent gas diffusion from the core to the confining fluid at high
sample pressures (Pan et al., 2010a).
Figure 4-3. Schematic plot of the tri-axial multi-gas rig.
Fig 1 Schematic plot of the Triaxial Multi-Gas Rig
Upstream
CH4
He
CO2
Downstreamm
CO2 Tube
Heater Upstream Cylinder Downstrea
Cylinder
Actuator
111.
122.
Vacuum
111.1 122.2
Pressure
Sensor
Pressure
Sensor
Pressure
Sensor
Pressure
Sensor
Pressure
Sensor
Pressure
Sensor
Pressure
Sensor
Upstream Injection
Pump
Confinement
Pump
Computer / DAQ
Water
Heater
Differential
Pressure
transducer
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-8
4.2.2 Permeability measurement
We use the pressure transient method to conduct the gas flow experiments in the low
permeability samples (Brace, et al., 1968; Hsieh et al., 1981), which has been widely used
due to its shorter test durations compared to steady state measurements. Permeability is
measured after the equilibrium state is reached for changes of either confining stress or
pore pressure. This method involves observing the decay of the differential pressure
between upstream and downstream cylinders across a sample of interest. The pressure
decay is combined with the cylinder volumes in the analysis to relate the flow through the
sample and thus determine the permeability (Brace et al., 1968; Hsieh et al., 1981;
Hildenbrand et al., 2002). The pressure decay curve can be modelled as:
tm
du
du ePP
PP
0,0,
(4-6)
where du PP is the pressure difference between the up and down stream cylinders used
for permeability tests, measured by a differential pressure transducer; 0,0, du PP is the
pressure difference between the up and down stream cylinders at initial stage (ranges from
60kPa to 100kPa for this work), t is the time and m is described below:
)11
(2
du
R
g VVV
LC
km
(4-7)
where k is permeability; gC is the gas compressibility; L is the sample length; RV is the
sample volume; uV and dV are the volume of the up and downstream cylinders.
Thus permeability can be obtained from Equation (4-7) after m is obtained from the
pressure decay curve. For this study, the approximate time allowed for each pressure decay
curve varies from several minutes to a few hours before pressure difference between the
upstream and downstream cylinders reach below 2.0 kPa.
4.2.3 Experimental procedure
Two Australian coal samples from the southern Sydney basin were used. The coal samples
were bituminous coal from the Bulli seam and cored to 4.50cm×10.55 cm (core: No.01)
and 4.55am×10.10 cm (core: No.02) in diameter and length, respectively.
The coal cores were first dried in a heated vacuum oven and the weight was measured.
Then the cores was installed in the sample cell and consolidated with a few load cycles
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-9
before carrying out experiments to make sure results were repeatable. The samples were
allowed to equilibrate for a few days inside the cell in the vacuum to remove the residual
gas and to reach the desired temperature.
Three gasses were used including helium, methane and carbon dioxide. Measurements were
conducted at 45°C and 35°C for core No.01 and No.02, respectively.
For core No.01 tests, permeability was firstly measured using helium at a constant pressure
difference (defined as confining stress minus pore pressure) of 3.0MPa. The injection
pressure started from 2.1MPa and increased by 2.0MPa for each step until reaching
10.1MPa. For methane and carbon dioxide injection, the pressure difference was kept at
2.0MPa during injection until the sorption reached equilibrium. The confining pressure was
changed to achieve the pressure differences of 4.0MPa and 6.0MPa, respectively. Four
different pore pressure steps were tested for both gasses. Permeability was measured after
each confining pressure or pore pressure change. After the measurements for each gas, coal
sample was vacuumed for a couple of days to remove the residual gas. The experimental
results for core No. 01 were also documented in Pan et al.'s work (2010a) and used for the
data analysis and modelling work presented in this paper.
For core No.02 tests, constant pressure differences with 2.0 and 3.0MPa were used for
helium injection. No CH4 flow-through experiments were conducted for this sample as
methane detector in the laboratory was out of order during the measurement so methane is
not allowed to use due to safety concerns. After the completion of helium test, the core
was vacuumed and CO2 was injected. For CO2 injection, the pressure difference was kept
at 2.0MPa until the pore pressure reached equilibrium. Then the confining pressure
changed to achieve the pressure difference of 3.0 and 4.0MPa. Three pore pressures of
CO2 injection were tested for this sample. Permeability is measured after each confining
pressure or pore pressure change for both helium and CO2.
4.2.4 Work flow of data analysis
The evaluation of coal effective stress coefficient and the elimination of effective stress
effect from sorption-induced strain effect are conducted through the following six
integrated steps: (1) measure permeability using helium gas under constant pressure
difference conditions; (2) calculate the effective stress coefficient; (3) calculate the
additional effective stress value and obtain the real effective stress value; (4) evaluate the
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-10
permeability change induced by the additional effective stress for adsorbing gas injection
under constant pressure difference conditions; (5) calibrate the permeability value to
eliminate the effective stress effect; and finally (6) obtain the sorption-induced strain effect
only on permeability change.
4.3 Results and discussion
4.3.1 Results for helium injection
Permeability measurements for the helium flow-though experiments are shown in Figure 4-
4. Permeability reductions with increasing pore pressure are observed for both cases under
the condition of constant pressure difference (defined as confining stress minus pore
pressure). Helium is almost a non-adsorbing gas to coal, thus the effect of matrix swelling
on permeability change is negligible. Therefore, the reduction in permeability can be
possibly attributed by two factors: (1) the Klinkenberg (1941) effect, especially in the low
pressure range; and (2) the effective stress effect. It has been observed that the Klinkenberg
effect diminishes with increase in gas pressure, because at high pressures (e.g. >2MPa) the
mean free path of the gas molecules (diameter is about 0.98 ˚A) is far less than the aperture
of the coal cleats (3–40 μm) (Laubach et al., 1998). Correspondingly, collisions between gas
molecules are more frequent than collisions between gas molecules and the solid walls
(Han et al., 2010). Therefore, the effective stress rather than the Klinkenberg effect is
considered as the cause for the permeability reduction at high pressures.
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-11
(a) Permeability vs. pore pressure with pressure difference for core No. 01.
(b) Permeability vs. pore pressure with pressure difference for core No. 02.
Figure 4-4. Permeability measured using helium.
Coal cleat compressibility calculation
In order to calculate the effective stress change, we firstly need to know how the effective
stress determines permeability change. An exponential functional form has been widely
y = 0.9922e-0.011x
R² = 0.9665
0.7
0.8
0.9
1
0 2 4 6 8 10 12
Pe
rme
ab
ilit
y (
md
)
Pore Pressure (MPa)
Pressure difference=3.0MPa
Flowing fluid: HeliumCore: No. 01
y = 1.1746e-0.056x
R² = 0.9887
y = 0.9366e-0.044x
R² = 0.9844
0.7
0.8
0.9
1
2 3 4 5 6 7
Pe
rme
ab
ilit
y v
alu
e, m
d
Pore pressure, MPa
Pressure difference=2.0MPa
Pressure difference=3.0MPa
Flowing fluid: HeliumCore: No. 02
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-12
used to calculate permeability with effective stress variation (Mckee et al., 1988; Seidle and
Huitt, 1995), and by combining with the influence of the effective stress coefficient on
effective stress change, the relation can be given as
pC fekk
3
0 (4-8)
where k is the permeability, 0k is the initial permeability and fC is the cleat
compressibility.
It should be noted that the original work from both Mckee et al. (1988) and Seidle and
Huitt (1995) have taken the effective stress coefficient to be unity, which means same
amount change of either pore pressure or confining stress has the same impact on effective
stress change.
From Equation (4-8) we can see that in order to calculate cleat compressibility value, the
following two steps are required: the pore pressure should be kept constant ( 0p ) and
only confining stress ( ) is allowed to change; permeability change with different
confining stress values should be measured. Then cleat compressibility values can be
obtained by regressing the permeability change with confining stress change
(
fCekk
3
0 ).
The experimental results for cleat compressibility in the presence of different gasses are
listed in Table 1.
Table 4-1. Compressibility values regarding to different gasses and pressures
Core Number Pore
pressure fC
Pore
pressure fC
Pore
pressure fC
(MPa) (MPa-1) (MPa) (MPa-1) (MPa) (MPa-1)
Helium CH4 CO2
No.01
(Pan et al., 2010a)
2.1 0.0848 0.9 0.0507 3.0 0.0606
10.1 0.0485 3.4 0.0472 6.4 0.0654
No.02
2.0 0.118 - - 1.3 0.142
3.0 0.102 - - 3.6 0.599
5.0 0.098 - - 5.0 0.687
The average compressibility values for helium, 0669.01 fC MPa-1 and 106.02 fC MPa-1
for cores No. 01 and 02, were used for the effective stress coefficient calculation.
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-13
Coal effective stress coefficient calculation:
Because the pressure difference was kept unchanged throughout the helium injection test
( p ), Equation (4-8) can be simplified to.
pC fekk
13
0 (4-9)
Substituting cleat compressibility values into the above equation, then the effective
coefficients (β) 0.945 for core No.01, and 0.842 and 0.855 for No.02 at pore pressures of
2.0 and 3.0MPa were obtained.
Therefore, the effective stress coefficient is not a constant ( 0.1 ) and it may increase
with increasing pore pressure. In this study, the incremental effective stress, due to not
unity effective stress coefficient ( 0.1 ), is called the additional effective stress, which
can be calculated by term p 1 . This stress should be added into the effective stress
term. In the following, the additional effective stress is recalculated for each case, and the
comparison for helium injection between original results and corrected magnitudes are
shown in Figure 4-5.
(a) Comparison of permeability behaviour for core No.01.
y = 1.3942e-0.146x
R² = 0.9845
y = 2.0956e-0.254x
R² = 0.9976
y = 2.1598e-0.255x
R² = 0.9977
y = 1.5127e-0.146x
R² = 0.9847
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
1.2
1.3
1.4
0 1 2 3 4 5 6 7
Pe
rme
ab
ilit
y (
md
)
Effective stress (MPa)
Pore Pressure=10.1MPa
Pore Pressure=2.1MPa
Corrected Value for 2.1 Mpa
Corrected Value for 10.1 Mpa
Flowing fluid: HeliumCore: No. 01
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-14
(b) Comparison of permeability behaviour for core No.02.
Figure 4-5. Comparison of permeability behavior between constant pressure difference
method and constant effective stress method for both cores.
It shows that the constant pressure difference method (original data) underestimates the
effective stress effect on permeability change. The underestimation becomes more
significant as the pore pressure increases.
In order to compare the difference between original results and corrected results, an
arbitrary error e is introduced and defined as
%100ty valuepermeabili Corrected
ty valuepermeabili Original-ty valuepermeabili Correctede (4-10)
For core No.01, the error changes from 3.44% at 2.1MPa to 9.57% at 10.1MPa. For core
No.02, the error is 13.7% for pore pressure at 2.0MPa. Hence the difference is significant
especially at higher pore pressure and for coal with a small effective stress coefficient.
4.3.2 Results for methane injection
It is reasonable to assume that the effective stress coefficient is constant for each test, as
the pressure difference was kept constant in our experiment. Based on the calculated
effective stress coefficient, β, the permeability data are corrected by subtracting the
additional effective stress induced permeability reduction from the original data. Equation
y = 2.5788e-0.355x
R² = 0.9933
y = 2.8898e-0.355x
R² = 0.9933
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
1.2
1.3
1.4
1.5 2 2.5 3 3.5 4 4.5
Pe
rme
ab
ilit
y (
md
)
Effective stress (MPa)
Flow fluid: HeliumCore: No.02
Pore Pressure=2.0MPa
Corrected valuel for 2.0MPa
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-15
(4-9) is used to calculate this additional permeability change. The initial permeability for
each case is calculated from the calibrated expression for helium, as shown in Figure 4-5.
With methane injection, the comparison between original permeability data and the
corrected ones for core No.01 is shown in Figure 4-6.
(a) Comparison of permeabilities for a pressure difference of 2.0MPa.
(b) Comparison of permeabilities for a pressure difference of 4.0MPa.
y = 0.8693e-0.06x
R² = 0.9984
y = 0.8513e-0.037x
R² = 0.9873
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
0 2 4 6 8 10 12 14
Pe
rme
ab
ilit
y (
md
)
Pore Pressure (MPa)
Pressure difference=2.0MPa
Corrected permeability value(2.0MPa)
Flowing fluid: MethaneCore: No. 01
y = 0.6172e-0.048x
R² = 0.9879
y = 0.6083e-0.026x
R² = 0.9324
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
0 5 10 15
Pe
rme
ab
ilit
y (
md
)
Pore Pressure (MPa)
Pressure difference=4.0MPa
Corrected permeability value (4.0MPa)
Flowing fluid: MethaneCore: No. 01
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-16
(c) Comparison of permeabilities for a pressure difference of 6.0MPa.
Figure 4-6. Comparison of permeabilities recovered by the constant pressure difference
method and constant effective stress method for CH4 injection for Core No.01.
4.3.3 Results for carbon dioxide injection
Following the same procedure as for the methane injection data, the comparison between
the uncorrected permeability data and corrected data for core No.01 and core No.02 using
CO2 are shown in Figures 4-7 and 4-8, respectively. Core No.01 involves four different
pore pressure steps, namely 3.0, 6.4, 9.8 and 13.3 MPa, and core No.02 also involves three
different pore pressure steps, namely 1.3, 3.6 and 5.0MPa.
y = 0.4636e-0.046x
R² = 0.9911
y = 0.4573e-0.024x
R² = 0.9469
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
0 5 10 15
Pe
rme
ab
ilit
y (
md
)
Pore Pressure (MPa)
Pressure difference=6.0MPa
Corrected permeability value (6.0MPa)
Flowing fluid: MethaneCore: No. 01
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-17
(a) Comparison of permeabilities for a pressure difference of 2.0MPa.
(b) Comparison of permeabilities for a pressure difference of 4.0MPa.
y = 0.7686e-0.087x
R² = 0.9941
y = 0.7053e-0.049x
R² = 0.9987
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0 5 10 15
Pe
rme
ab
ilit
y (
md
)
Pore Pressure (MPa)
Pressure difference=2.0MPa
Corrected permeability value (2.0MPa)
Flowing fluid: CO2Core: No. 01
y = 0.5665e-0.109x
R² = 0.9791
y = 0.4956e-0.054x
R² = 0.9485
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0 5 10 15
Pe
rme
ab
ilit
y (
md
)
Pore Pressure (MPa)
Pressure difference=4.0MPa
Corrected permeability value (4.0MPa)
Flowing fluid: CO2Core: No. 01
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-18
(c) Comparison of permeabilities for a pressure difference of 6.0MPa.
Figure 4-7. Comparison of permeabilities between the constant pressure difference method
and the constant effective stress method for CO2 injection for core No.01.
(a) Comparison of permeabilities for a pressure difference of 2.0MPa.
y = 0.4999e-0.163x
R² = 0.9728
y = 0.3822e-0.078x
R² = 0.9461
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0 5 10 15
Pe
rme
ab
ilit
y (
md
)
Pore Pressure (MPa)
Pressure difference=6.0MPa
Corrected permeability value (6.0MPa)
Flowing fluid: CO2Core: No. 01
y = 1.1421e-0.451x
R² = 0.9569
y = 0.8036e-0.14x
R² = 0.9981
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0 1 2 3 4 5 6
Pe
rme
ab
ilit
y (
md
)
Pore Pressure (MPa)
Pressure difference=2.0MPa
Corrected value for 2.0MPa
Flowing fluid: CO2Core: No. 02
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-19
(b) Comparison of permeabilities for a pressure difference of 3.0MPa.
(c) Comparison of permeabilities for a pressure difference of 4.0MPa.
Figure 4-8. Comparison of permeabilities obtained between the constant pressure
difference method and the constant effective stress method for CO2 injection for core
No.02.
4.4 Discussion
From Figures 4-6 to 4-8, we can see that generally the original data follow an exponential
function better than the corrected one. The original data includes the effects from two
factors: effective stress and sorption-induced strain, while the corrected data includes the
y = 1.6564e-0.945x
R² = 0.8936
y = 0.5445e-0.218x
R² = 0.9956
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0 1 2 3 4 5 6
Pe
rme
ab
ilit
y (
md
)
Pore Pressure (MPa)
Pressure difference=3.0MPa
Corrected value for 3.0MPa
Flowing fluid: CO2Core: No. 02
y = 1.1153e-0.973x
R² = 0.9434
y = 0.388e-0.275x
R² = 0.9631
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0 1 2 3 4 5 6
Pe
rme
ab
ilit
y (
md
)
Pore Pressure (MPa)
Pressure difference=4.0MPa
Corrected value for 4.0MPa
Flowing fluid: CO2Core: No. 02
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-20
sorption-induced strain only, which follows a Langmuir like equation. This comparison
supports that the corrected results represent the impact from sorption-induced strain only.
A summary of the permeability errors, as defined in Equation (4-10), for both cores No.01
and 02 are listed in Tables 4-2 and 4-3. For both coal samples, a significant difference of
permeability change vs. pore pressure is obtained between original results (β=1.0) and
corrected results. This demonstrates that the constant pressure difference method
overestimates the sorption-induced permeability change especially in the high pore pressure
range. This is because part of effective stress induced permeability reduction is considered
to be from sorption-induced strain. The overestimation is enhanced as the pore pressure
increases. For core No.01, an error as high as 27.86% is observed with CH4 injection within
the pressure range of the experiments. This error may be as large as 60% with CO2
injection. For core No.02, it is even higher than for core No.01 with CO2 injection. This
indicates that it is necessary to correct the permeability data by eliminating the influence of
the additional effective stress from the swelling strain.
Table 4-2. Summary of permeability errors under different pressure difference values with
CH4 injection
Pore pressure
(MPa)
Error (2.0MPa)
(%)
Error (4.0MPa)
(%)
Error (6.0MPa)
(%)
Core No.01
0.92 1.18 1.56 1.39
3.54 5.20 6.97 5.79
7.41 12.18 15.58 13.41
12.80 24.33 27.86 24.08
Table 4-3. Summary of permeability errors under different pressure difference values for
CO2 injection
Pore pressure
(MPa)
Error (2.0MPa)
(%)
Error (4.0MPa)
(%)
Error (6.0MPa)
(%)
Core
No.01
3.0 5.36 6.70 6.94
6.4 13.02 15.80 17.39
9.8 23.44 32.12 40.76
13.3 36.28 46.03 61.04
Pore pressure
(MPa)
Error (2.0MPa)
(%)
Error (3.0MPa)
(%)
Error (4.0MPa)
(%)
Core 1.3 12.72 12.19 9.93
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-21
No.02 3.6 43.74 54.09 57.03
5.0 73.61 94.96 94.06
In the following part, comparison of permeability change between sorption-induced and
effective stress induced is conducted to evaluate the contribution from each process.
Based on the corrected permeability data, the absolute permeability reduction value, rk ,
induced by sorption strain with different gas species is calculated to compare the
contribution of each component on permeability change, i.e. due to CO2 and CH4. rk is
defined as
fmmr kkk 0 (4-11)
where mk0
is the initial permeability at effective stress m with zero pore pressure, and fmk
represents permeability value at the final experimental pore pressure (e.g. 13.3MPa for core
No.01 and 5.0MPa No.02) at effective stress m .
The comparison is shown in Figure 4-9. This figure demonstrates that, for core No.01,
CO2 injection can cause about a 1.5 times larger permeability reduction than CH4 under the
same pressure conditions, and about 2.1 times larger than effective stress induced
permeability reduction using Helium. Similar phenomena for core No.02 are also observed,
showing CO2 could cause a 1.9 times larger permeability reduction than effective stress
induced permeability reduction using helium. Note that the first point for effective stress
induced permeability change curve of core No.02 is quite different from other points; this
may be due to the difficulty to estimate the initial coal permeability value at atmosphere
condition, which is regressed from other pore pressure measurements in this study.
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-22
Figure 4-9. Comparison of permeability change caused by compaction and swelling strain.
This set of tests shows that the effective stress induced permeability reduction is close to
that induced by CH4 injection, and the difference to CO2 injection is not as significant as
field observations. Field trails have shown dramatic permeability reduction/increase with
adsorbing gas injection/production (Fujioka et al., 2010; Mavor and Vaughn, 1998; Mavor
et al., 2004; Palmer, 2009; van Bergen et al., 2009a,b; Wong et l., 2007). This difference
could be attributed to the difference between the laboratory experimental conditions and in
situ conditions. Under in situ conditions the effect is superimposed upon the response
which is less likely in a small coal sample tested in the laboratory. The sorption-induced
strain may close these natural cleats, and in turn causes dramatic changes in permeability.
However, under experimental conditions, the cores are normally consolidated before the
experiments, which artificially close the fractures and in turn diminishes the sorption-
induced strain effects.
4.5 Conclusions
This study has demonstrated that the effective stress coefficient for coal is not equal to
unity and may be effective stress dependent. This conclusion is derived based on the
experimental observations of a series of gas flow-through experiments all under the
conditions of a constant pressure difference between confining stress and pore pressure.
First, the effective stress coefficient is obtained by the non-adsorbing gas (helium) flow-
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1 2 3 4 5 6 7
Pe
rme
ab
ilit
y r
ed
uc
tio
n (
md
)
Pressure difference (MPa)
CO2 (Core No.02)
CO2(Core No.01)
CH4 (Core No.01)
Effective stress induced (Core No.01)
Effective stress induced (Core No.02)
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-23
though experiments. In these experiments, the impact of gas sorption is negligible and any
permeability change with pore pressure is considered to be due to the variation in the
effective stress coefficient. Second, the change in permeability resulting from the non-
adsorbing gas experiments is used to calibrate the subsequent experiments using adsorbing
gasses (carbon dioxide and methane) where the sole effect of sorption-induced strain on
permeability change is obtained. This finding may be more important because even though
coal is a weak rock, the grain compressibility is larger than expected, especially at high pore
pressures. Comparison between measured permeability data and calibrated results
demonstrates that the effective stress coefficient could play an important role in the
evaluation of permeability change in adsorbing gasses.
4.6 Acknowledgments
This work was supported by WA:ERA, the Western Australia CSIRO-University
Postgraduate Research Scholarship, National Research Flagship Energy Transformed Top-
up Scholarship, and by NIOSH under contract 200-2008-25702. These supports are
gratefully acknowledged. We also wish to express our sincere thanks to Mr. Michael
Camilleri, for his assistance and guidance in carrying out these tests in this work.
4.7 References
Alam, M.M., Borre, M.K., Fabricius, I.L., Hedegaard, K., Røgen, B., Hossain, Z., Krogsbøll, A.S., 2010. Biot's
coefficient as an indicator of strength and porosity reduction: calcareous sediments from Kerguelen
Plateau. Journal of Petroleum Science and Engineering 70, 282-297.
Bear, J., 1972. Dynamics of Fluids in Porous Media. American Elsevier Pub. Co., New York.
Bernabe, Y., 1987. The effective pressure law for permeability during pore pressure and confining pressure
cycling of several crystalline rocks. Journal. Geophysical. Research 92, 649-657.
Biot, M.A., 1941. General theory of three-dimensional consolidation. Journal of Applied Physics 12, 155-164.
Brace, W.F., Walsh, J.B., Frangos, W.T., 1968. Permeability of granite under high pressure. Journal.
Geophysical. Research 73, 2225-2236.
Bustin, R.M., Cui, X., Chikatamarla, L., 2008. Impacts of volumetric strain on CO2 sequestration in coals and
enhanced CH4 recovery. AAPG Bulletin 92, 15-29.
Chikatamarla, L., Cui, X., Bustin, R.M., 2004. Implications of volumetric swelling/shrinkage of coal in
sequestration of acid gasses, International Coalbed Methane Symposium, Tuscaloosa, Alabama, Paper
No. 0435.
Clarkson, C.R., Bustin, R.M., 2000. Binary gas adsorption/desorption isotherms: effect of moisture and coal
composition upon carbon dioxide selectivity over methane. International Journal of Coal Geology 42,
241-271.
Cui, X., Bustin, R.M., Chikatamarla, L., 2007. Adsorption-induced coal swelling and stress: implications for
methane production and acid gas sequestration into coal seams. Journal. Geophysical. Research 112,
B10202.
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-24
Day, S., Fry, R., Sakurovs, R., 2008. Swelling of Australian coals in supercritical CO2. International Journal of
Coal Geology 74, 41-52.
Fujioka, M., Yamaguchi, S., Nako, M., 2010. CO2-ECBM field tests in the Ishikari Coal Basin of Japan.
International Journal of Coal Geology 82, 287-298.
Gash, B.W., 1991. Measurement of "Rock Properties" in Coal for Coalbed Methane Production, SPE Annual
Technical Conference and Exhibition. Society of Petroleum Engineers, Inc., Dallas, TX.
Ghabezloo, S., Sulem, J., Guédon, S., Martineau, F., 2009. Effective stress law for the permeability of a
limestone. International Journal of Rock Mechanics and Mining Sciences 46, 297-306.
Han, F., Busch, A., van Wageningen, N., Yang, J., Liu, Z., Krooss, B.M., 2010. Experimental study of gas and
water transport processes in the inter-cleat (matrix) system of coal: anthracite from Qinshui Basin,
China. International Journal of Coal Geology 81, 128-138.
Harpalani, S., Chen, G., 1997. Influence of gas production induced volumetric strain on permeability of coal.
Geotechnical and Geological Engineering 15, 303-325.
Harpalani, S., Schraufnagel, R.A., 1990. Shrinkage of coal matrix with release of gas and its impact on
permeability of coal. Fuel 69, 551-556.
Hildenbrand, A., Schlömer, S., Krooss, B.M., 2002. Gas breakthrough experiments on fine-grained
sedimentary rocks. Geofluids 2, 3-23.
Hsieh, P.A., Tracy, J.V., Neuzil, C.E., Bredehoeft, J.D., Silliman, S.E., 1981. A transient laboratory method
for determining the hydraulic properties of 'tight' rocks--I. Theory. International Journal of Rock
Mechanics and Mining Sciences & Geomechanics Abstracts 18, 245-252.
Huy, P.Q., Sasaki, K., Sugai, Y., Ichikawa, S., 2010. Carbon dioxide gas permeability of coal core samples and
estimation of fracture aperture width. International Journal of Coal Geology 83, 1-10.
Karacan, C.Ö., 2003. Heterogeneous sorption and swelling in a confined and stressed coal during CO2
Injection. Energy & Fuels 17, 1595-1608.
Klinkenberg, L.J., 1941. The Permeability Of Porous Media To Liquids And Gasses, Drilling and Production
Practice. American Petroleum Institute.
Laubach, S.E., Marrett, R.A., Olson, J.E., Scott, A.R., 1998. Characteristics and origins of coal cleat: a review.
International Journal of Coal Geology 35, 175-207.
Levine, J.R., 1996. Model study of the influence of matrix shrinkage on absolute permeability of coal bed
reservoirs, in: Gayer, R., Harris, I. (Eds.), Coalbed Methane and Coal Geology. Geological Society
Special Publication No 109. London, pp. 197- 212.
Lin, W., Tang, G.-Q., Kovscek, A.R., 2008. Sorption-induced permeability change of coal during gas-injection
processes. SPE Reservoir Evaluation & Engineering 11.
Liu, C.J., Wang, G.X., Sang, S.X., Rudolph, V., 2010a. Changes in pore structure of anthracite coal associated
with CO2 sequestration process. Fuel 89, 2665-2672.
Liu, J., Chen, Z., Elsworth, D., Miao, X., Mao, X., 2010b. Evaluation of stress-controlled coal swelling
processes. International Journal of Coal Geology DOI: 10.1016/j.coal.2010.06.005.
Mavor, M.J., Gunter, W.D., 2006. Secondary porosity and permeability of coal vs. gas composition and
pressure. SPE Reservoir Evaluation & Engineering 9, 114-125.
Mavor, M.J., Gunter, W.D., Robinson, J.R., 2004. Alberta multiwell micro-pilot testing for CBM properties,
enhanced methane recovery and CO2 storage potential, SPE Annual Technical Conference and
Exhibition, Houston, Texas.
Mavor, M.J., Vaughn, J.E., 1998. Increasing coal absolute permeability in the San Juan Basin Fruitland
Formation. SPE Reservoir Evaluation & Engineering 1, 201-206.
Mazumder, S., Karnik, A.A., Wolf, K.-H.A.A., 2006. Swelling of coal in response to CO2 sequestration for
ECBM and its effect on fracture permeability. SPE Journal 11, pp. 390-398.
Mazumder, S., Wolf, K.H., 2008. Differential swelling and permeability change of coal in response to CO2
injection for ECBM. International Journal of Coal Geology 74, 123-138.
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-25
McKee, C.R., Bumb, A.C., Koenig, R.A., 1988. Stress-dependent permeability and porosity of coal and other
geologic formations. SPE Formation Evaluation 3, 81-91.
Moffat, D.H., Weale, K.E., 1955. Sorption by coal of methane at high pressures. Fuel 34, 449-462.
Nur, A., Byerlee, J.D., 1971. An exact effective stress law for elastic deformation of rock with fluids. J.
Geophys. Res. 76, 6414-6419.
Ozdemir, E., Schroeder, K., 2009. Effect of moisture on adsorption isotherms and adsorption capacities of
CO2 on coals. Energy & Fuels 23, 2821-2831.
Palmer, I., 2009. Permeability changes in coal: Analytical modeling. International Journal of Coal Geology 77,
119-126.
Palmer, I., Mansoori, J., 1996. How permeability depends on stress and pore pressure in coalbeds: a new
model, SPE Annual Technical Conference and Exhibition. 1996 Copyright 1996, Society of Petroleum
Engineers, Inc., Denver, Colorado.
Pan, Z., Connell, L.D., 2007. A theoretical model for gas adsorption-induced coal swelling. International
Journal of Coal Geology 69, 243-252.
Pan, Z., Connell, L.D., 2011. Modelling of anisotropic coal swelling and its impact on permeability behaviour
for primary and enhanced coalbed methane recovery. International Journal of Coal Geology 85, 257-
267.
Pan, Z., Connell, L.D., Camilleri, M., 2010a. Laboratory characterisation of coal reservoir permeability for
primary and enhanced coalbed methane recovery. International Journal of Coal Geology 82, 252-261.
Pan, Z., Connell, L.D., Camilleri, M., Connelly, L., 2010b. Effects of matrix moisture on gas diffusion and
flow in coal. Fuel 89, 3207-3217.
Patching, T.H., 1965. Variations in permeability of coal, Proc. 3rd Rock Mech. Symp, Univ. of Toronto, U.K.,
pp. 185-199.
Pini, R., Ottiger, S., Burlini, L., Storti, G., Mazzotti, M., 2009. Role of adsorption and swelling on the
dynamics of gas injection in coal. J. Geophys. Res. 114, B04203.
Reucroft, P.J., Patel, H., 1986. Gas-induced swelling in coal. Fuel 65, 816-820.
Reucroft, P.J., Sethuraman, A.R., 1987. Effect of pressure on carbon dioxide induced coal swelling. Energy &
Fuels 1, 72-75.
Robertson, E.P., Christiansen, R.L., 2005. Measurement of sorption-induced strain., International Coalbed
Methane Symposium, Tuscaloosa, Alabama,17- 19 May, Paper 0532.
Robertson, E.P., Christiansen, R.L., 2007. Modeling laboratory permeability in coal using sorption-induced
strain data. SPE Reservoir Evaluation & Engineering 10, pp. 260-269.
Robin, P.-Y.F., 1973. Note on effective pressure. Journal Geophysical Research 78, 2434-2437.
Seidle, J.R., Huitt, L.G., 1995. Experimental measurement of coal matrix shrinkage due to gas desorption and
implications for cleat permeability increases, International Meeting on Petroleum Engineering. 1995
Copyright 1995, Society of Petroleum Engineers, Inc., Beijing, China.
Seidle, J.P., Jeansonne, M.W., Erickson, D.J., 1992. Application of matchstick geometry to stress dependent
permeability in coals. SPE Rocky Mountain Regional Meeting, SPE 24361, Casper, Wyoming.
Shi, J.Q., Durucan, S., 2004. Drawdown induced changes in permeability of coalbeds: a new interpretation of
the reservoir response to primary recovery. Transport in Porous Media 56, 1-16.
Siriwardane, H., Haljasmaa, I., McLendon, R., Irdi, G., Soong, Y., Bromhal, G., 2009. Influence of carbon
dioxide on coal permeability determined by pressure transient methods. International Journal of Coal
Geology 77, 109-118.
Somerton, W.H., Söylemezoglu, I.M., Dudley, R.C., 1975. Effect of stress on permeability of coal.
International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts 12, 129-
145.
St. George, J.D., Barakat, M.A., 2001. The change in effective stress associated with shrinkage from gas
desorption in coal. International Journal of Coal Geology 45, 105-113.
School of Mechanical and Chemical Engineering
Chapter 4 The University of Western Australia
4-26
Terzaghi, K.V., 1923. Die Berechnung der durchassigkeitsziffer des Tones aus dem Verlauf der
hydrodynamischen Spannungserscheinungen. Sitzungbericht Akademie Wissenschaft Wien,
Mathematik Naturwissenschaft 132, 105-126.
van Bergen, F., Krzystolik, P., van Wageningen, N., Pagnier, H., Jura, B., Skiba, J., Winthaegen, P., Kobiela,
Z., 2009. Production of gas from coal seams in the Upper Silesian Coal Basin in Poland in the post-
injection period of an ECBM pilot site. International Journal of Coal Geology 77, 175-187.
van Bergen, F., Spiers, C., Floor, G., Bots, P., 2009. Strain development in unconfined coals exposed to CO2,
CH4 and Ar: Effect of moisture. International Journal of Coal Geology 77, 43-53.
Walsh, J.B., 1981. Effect of pore pressure and confining pressure on fracture permeability. International
Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts 18, 429-435.
Wang, G.X., Wei, X.R., Wang, K., Massarotto, P., Rudolph, V., 2010. Sorption-induced swelling/shrinkage
and permeability of coal under stressed adsorption/desorption conditions. International Journal of
Coal Geology 83, 46-54.
White, C.M., Smith, D.H., Jones, K.L., Goodman, A.L., Jikich, S.A., LaCount, R.B., DuBose, S.B., Ozdemir,
E., Morsi, B.I., Schroeder, K.T., 2005. Sequestration of carbon dioxide in coal with enhanced coalbed
methane recovery: a Review. Energy & Fuels 19, 659-724.
Wong, S., Law, D., Deng, X., Robinson, J., Kadatz, B., Gunter, W.D., Jianping, Y., Sanli, F., Zhiqiang, F.,
2007. Enhanced coalbed methane and CO2 storage in anthracitic coals–micro-pilot test at South
Qinshui, Shanxi, China. International Journal of Greenhouse Gas Control 1, 215-222.
Yangsheng, Z., Yaoqing, H., Jingping, W., Dong, Y., 2003. The Experimental Approach to Effective Stress
Law of Coal Mass by Effect of Methane. Transport in Porous Media 53, 235-244.
Zhang, K., Zhou, H., Hu, D., Zhao, Y., Feng, X., 2009. Theoretical model of effective stress coefficient for
rock/soil-like porous materials. Acta Mechanica Solida Sinica 22, 251-260.
Zheng, Z., 1993. Compressibility of porous rocks under different stress conditions, The 34th U.S. Symposium
on Rock Mechanics (USRMS). Geological Engineering Program, UW-Madison. Permission to
Distribute - American Rock Mechanics Association, Madison, WI.
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-1
Chapter 5
Influence of the effective stress coefficient and sorption-induced strain on the evolution of coal permeability: Model development and analysis
Zhongwei Chena,b, Jishan Liua, Zhejun Panb, Luke D. Connellb, Derek Elsworthc
a School of Mechanical and Chemical Engineering, The University of Western Australia, WA 6009, Australia
b CSIRO Earth Science and Resource Engineering, Private Bag 10, Clayton South, Victoria 3169, Australia
c Department of Energy and Mineral Engineering, Penn State University, PA 16802-5000, USA
ABSTRACT: A series of coal permeability experiments was conducted for coal samples
infiltrated both with non-adsorbing and adsorbing gasses - all under conditions of constant
pressure difference between the confining stress and the pore pressure. The experimental
results show that even under controlled stress conditions, coal permeability decreases with
respect to pore pressure during the injection of adsorbing gasses. This conclusion is
apparently not congruent with our conceptual understanding: when coal samples are free to
swell/shrink then no effect of swelling/shrinkage strain should be apparent on the
permeability under controlled stress conditions. In this study, we developed a
phenomenological permeability model to explain this enigmatic behavior of coal
permeability evolution under the influence of gas sorption by combining the effect of
swelling strain with that of the mechanical effective stress. For the mechanical effective
stress effect, we use the concept of natural strain to define its impact on the change in
fracture aperture; for the swelling strain effect, we introduce a partition ratio to define the
contribution of swelling strain to the fracture aperture reduction. The resulting coal
permeability model is defined as a function of both the effective stress and the swelling
strain. Compared to other commonly used models under specific boundary conditions,
such as Palmer-Mansoori (P-M), Shi-Durucan (S-D) and Cui-Bustin (C-B) models, our
model results match the experimental measurements quite well. We match the experimental
data with the model results for the correct reason, i.e., the model conditions are consistent
with the experimental conditions (both are stress-controlled), while other models only
match the data for a different reason (the model condition is uniaxial strain but the
experimental condition is stress-controlled). We have also implemented our permeability
model into a fully coupled coal deformation and gas transport finite element model to
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-2
recover the important non-linear responses due to the effective stress effects where
mechanical influences are rigorously coupled with the gas transport system.
Keywords: coal permeability; swelling strain; effective stress effect; CO2 storage
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-3
5.1 Introduction
Coal Bed Methane (CBM) is naturally occurring methane gas (CH4) in coal seams. Methane
was long considered a major problem in underground coal mining but now CBM is
recognized as a valuable resource. Australia has vast reserves of coal-bed methane (about
310 to 410 trillion m3) (White et al., 2005) and has attracted billions of dollars in foreign
investment to develop this resource. CBM recovery triggers a series of coal-gas interactions.
For gas production, the reduction of gas pressure increases effective stress which in turn
closes fracture aperture and reduces the permeability (McKee et al., 1988; Seidle and Huitt,
1995; Palmer and Mansoori, 1996). As gas pressure reduces below the desorption point,
methane is released from coal matrix to the fracture network and coal matrix shrinks. As a
direct consequence of this matrix shrinkage the fractures dilate and fracture permeability
correspondingly increases (Harpalani and Schraufnagel, 1990). Thus a rapid initial
reduction in fracture permeability (due to change in effective stress) is supplanted by a slow
increase in permeability (with matrix shrinkage). Whether the ultimate, long-term,
permeability is greater or less than the initial permeability depends on the net influence of
these dual competing mechanisms (Shi and Durucan, 2004; Chen et al., 2008; Connell,
2009). Therefore, understanding the transient characteristics of permeability evolution in
fractured coals is of fundamental importance to the CBM recovery and CO2 storage in coal,
which has dual and complementary benefits: the enhanced production of methane and
concurrent long-term storage of CO2.
A broad variety of models have evolved to represent the effects of sorption, swelling and
effective stresses on the dynamic evolution of permeability over last few decades. In the
latest review (Liu et al., 2011), these models are classified into two groups: permeability
models under conditions of uniaxial strain and permeability models under conditions of
variable stress.
Somerton et al. (1975) investigated the permeability of fractured coal to methane and
presented a correlation equation in the prediction of permeability with mean stress. Gray
(1987) considered the changes in the cleat permeability as a function of the prevailing
effective horizontal stresses, and firstly incorporated the influence of matrix shrinkage into
a permeability model. Seidle and Huitt (1995) developed a conceptual matchstick model to
explain coal permeability decrease with increasing effective stress. Other stress-based coal
permeability models include Harpalani and Chen (1997), Gilman and Beckie (2000), Shi
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-4
and Durucan (S-D) (2004), Cui and Bustin (C-B) (2005). Based on cubic geometry,
Robertson-Christiansen (2006) described the derivation of a new equation that can be used
to model the permeability behaviour of a fractured, sorptive-elastic medium, such as coal,
under variable stress conditions. Ma et al. (2010) proposed a permeability model based on
the volumetric balance between the bulk coal, solid grains and pores, using the constant
volume theory proposed by Massarotto et al. (2009).
A number of coal permeability models were developed based on strains. Mckee et al. (1988)
developed a theoretical permeability model using matrix compressibility as a fundamental
property, but did not include the effect of sorption-induced strain on permeability change.
Sawyer et al. (1990) proposed a permeability model assuming that fracture porosity (to
which permeability can be directly related) is a linear function of changes in gas pressure
and concentration. Palmer and Mansoori (P-M) (1996) presented a theoretical model for
calculating pore volume compressibility and permeability in coals as a function of effective
stress and matrix shrinkage. The P-M model was updated in Palmer et al. (2007). Similarly,
the Advanced Resources International (ARI) group developed another permeability model
(Pekot and Reeves, 2002). This model does not have a geomechanics framework, but
instead extracts matrix strain changes from a Langmuir curve type of strain versus reservoir
pressure, which is assumed to be proportional to the gas concentration curve. Zhang et al.
(2008) developed a permeability model under variable stress conditions, and was extended
to CO2-ECBM conditions by Chen et al. (2009; 2010). Connell et al. (2010) presented two
analytical permeability models for tri-axial strain and stress conditions.
Pan and Connell (2007) developed a theoretical model for sorption-induced strain and
applied to single-component adsorption/strain experimental data. Clarkson (2008)
expanded this theoretical model to calculate the sorption-strain component of the P-M
model. Pan and Connell (2011a) developed an anisotropic swelling model based on their
swelling model (Pan and Connell, 2007). The dependence of coal permeability on pore
volume compressibility was also investigated (Shi and Durucan, 2010; Tonnsen and
Miskimins, 2010).
As reviewed above, there are a large collection of coal permeability models from empirical
ones to theoretical ones. These models normally have a set of common assumptions: (1)
the overburden stress remains constant; (2) coal deforms under the uniaxial strain
condition; (3) the effective stress coefficient is assumed as one; and (4) the sorption-
induced strain is totally counteracted by the closure of the fracture aperture. These
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-5
assumptions have limited their applicability as Liu et al. (2011) concluded that current
models have so far failed to explain the results from stress-controlled shrinkage/swelling
laboratory tests and have only achieved some limited success in explaining and matching in
situ data. Liu et al. (2011) considered the main reason for these failures is the impact of coal
matrix-fracture compartment interactions has not yet been understood well and further
improvements are necessary as demonstrated in latest studies (Connell et al., 2010; Liu and
Rutqvist, 2010; Izadi et al., 2011). In this study, a coal permeability model based on coal
matrix-fracture interaction was developed and then implemented into a fully coupled coal
deformation and gas transport finite element model to recover the important non-linear
responses due to the effective stress effects.
5.2 Permeability model development
Previous work of Chen et al. (2011) has reported the findings of a series of experiments
conducted for coal samples infiltrated both with non-adsorbing and adsorbing gasses - all
under conditions of constant pressure difference between the confining stress and the pore
pressure. Observations have demonstrated that even under controlled stress conditions the
injection of adsorbing gasses actually does reduce coal permeability. The swelling strain
effect has also been separated from the effective stress effect. In this section, we combined
the swelling strain effect with the mechanical effective stress effect into a
phenomenological permeability model to explain the enigmatic behaviour of coal
permeability evolution under the influence of gas sorption.
Experimental observations have shown that swelling response to the infiltration of CO2
exhibits two key features: (1) as CO2 infiltrates coal fracture, coal matrix swells and
permeability generally reduces if the gas pressure is not very high. This occurs regardless of
the mechanical constraint applied to the cracked coal sample; (2) the permeability recovers
with increasing gas pressure as effective stress effects dominate in the absence of swelling-
induced closure.
We use the idealized model as illustrated in Figure 5-1 to represent a single fracture within
a representative elementary volume. This representation is through two steps: the effective
stress is applied first in a non-adsorbing medium as shown in Figure 5-1(a) and then only
the pore pressure in an adsorbing medium as shown in Figure 5-1(b).
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-6
(a) (b)
Figure 5-1. Illustration of two-step loading process: (1) effective stress is applied in a non-
adsorbing medium; (2) pore pressure is applied in an adsorbing medium. (a) Effective
stress effects; (b) Swelling strain effects. p is pore pressure, is overburden stress, b
is fracture aperture and s is sorption-induced bulk dimension change.
5.2.1 Evaluation of effective stress effects
In this section, we define the behaviour of coal fracture where the effective stress is applied
in a non-adsorbing medium. Coals are viewed as naturally fractured reservoirs with a matrix
that is often assumed to have a negligible permeability in comparison to the fracture system.
These fractures in coal are known as cleats with the cleat aperture sensitive to the effective
stress, and increased effective stress acting to decrease the cleat aperture and thus
permeability.
A number of empirical and theoretical expressions exist in the literature for describing the
observed relationship between effective stress and fracture aperture (Daley, 2006; Liu et al.,
2009).
Liu et al. (2009) argued that porous and fractured rock (or coal) is inherently heterogeneous
and includes both a solid phase and pores (or fractures). Thus, an accurate description of
the deformation of the rock (or coal) is important for coupled mechanical and hydrological
processes, because fluid flow occurs in pores and fractures.
To deal with this issue, it is conceptualized that the fracture system has two parts, which
are subject to the same stress, but follow different varieties of Hooke's law: the hard part
follows the engineering-strain based Hooke's law, and the soft part obeys the natural-strain
∆s
+
σ
σ
p
p
<b><b>p p p
p pp
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-7
based Hooke's law, as shown in Figure 5-2. This treatment is consistent with previous
studies (Mavko and Jizba, 1991; Berryman, 2006; Liu et al., 2009).
Figure 5-2. Conceptualization of fracture system with the hard and soft parts. They follow
engineering-strain based and natural-strain based Hooke's law, respectively. 0b is the totally
unstressed fracture aperture, eb ,0and
tb ,0 are the unstressed fracture apertures for the hard
part and the soft part, respectively.
Considering a fracture to be embedded into a core sample subject to a stress, , Hooke's
law for the hard part can be expressed as
ebe dKd , (5-1)
where eK is the bulk modulus for the hard part of fracture system, and
eb, is the
engineering strain of fracture aperture (Jaeger et al., 2007), defined as
e
eeb
b
dbd
,0
, (5-2)
where eb ,0 is the unstressed fracture aperture for the hard part.
For the soft part, the following relation is used
tbt dKd , (5-3)
wheretK is the bulk modulus for the soft part of fracture system, and
tbd , is the true or
natural strain, defined as
t
ttb
b
dbd , (5-4)
b0,t
Soft Part
b0,e
Hard Part
b0
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-8
where tb is the fracture aperture for soft part under the current state of stress. Subscripts e
and t (for ―engineering‖ and ―true‖, respectively) refer to the ―hard‖ and ―soft‖ parts in
coal fracture system.
Freed (1995) provided a historical review of the development of the concept of natural
strain and argued that the natural strain should be used for accurately describing material
deformation.
Using the condition that tt bb ,0 and
ee bb ,0 for 0 , the engineering strain and natural
strain can be integrated into the following expressions
e
eee
Kbb
1,0 (5-5)
t
ett
Kbb
exp,0 (5-6)
Based on the above analysis, the total fracture aperture, stress
b , under stressed conditions
can be given as
testressbbb (5-7)
Combining the relation of fracture aperture change for both ―hard‖ and ―soft‖ parts, as
shown in Equations (5-5) and (5-6), yields the fracture aperture change with effective stress
t
et
e
eestress K
bK
bb
exp1 ,0,0 (5-8)
eK is generally several orders larger than tK . Therefore, the above equation can be
simplified into
t
etestress K
bbb
exp,0,0 (5-9)
From Figure 5-2 we can see that te bbb ,0,00 , thus the following expression can be
obtained.
)exp(,00,0 efeestressCbbbb (5-10)
where fC is coal fracture compressibility, defined as
tf KC 1 . 0b is the initial unstressed
total fracture aperture opening.
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-9
5.2.2 Evaluation of sorption-induced strain effects
We represent the behaviour of coal permeability where fluid pressures are applied in an
adsorbing medium. When the influence of swelling strain is investigated, a common way is
to assume that the swelling strain is totally accommodated by the closure of fracture
aperture, which could dramatically overestimate the influence of the swelling strain
(Robertson, 2005; Connell et al., 2010; Liu and Rutqvist, 2010). Based on the illustration
from Figure 5-1, we believe that only part of total swelling strain contributes to fracture
aperture change and the remaining portion of the swelling strain contributes to coal bulk
deformation, and a partition factor, f , is introduced to estimate this contribution, then
the fracture aperture change can be given as
3
s
swellingsfb
(5-11)
whereswelling
b is the fracture aperture change induced by the swelling strain only, and s is
the fracture spacing, f ranges from 0 to 1.0, s is the volumetric free swelling strain
change, which can be calculated as
LL
Lpp
p
pp
p
0
0 with Langmuir type, as shown
in Figure 5-3 (Pan et al., 2011b), and 31 ssf term accounts for coal bulk
deformation. L and Lp represent maximum volumetric swelling strain and Langmuir
pressure constant, respectively.
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-10
Figure 5-3. Volumetric swelling strain with Langmuir type function matching: experimental
data is shown with dot points, and solid line is the matching curve with Langmuir type
function.
The influence of swelling process on the internal stress distribution of coal fracture is
analysed in the following section. From Figure 5-1 we can see that after coal matrix swells,
the swelling strain increases the contact area of the cleat system, and in turn closes the
fracture. The cross-section of a representative element was shown in Figure 5-4, where the
contact area of the cleat system varies with the swelling process. Because the total stress (or
confining stress) along this section is kept constant during swelling process (second stage),
this internal stress along the cross section should always be balanced with surrounding
boundary. Therefore, the stress balance expression is
fefbeb ApAp 11 (5-12)
where eb and
ef are the equivalent internal effective stress before and after swelling
respectively, and bA and
fA are the effective contact areas before and after swelling,
respectively.
Because fb AA for the swelling case, the internal effective stress actually increases with
the increase of the contact area during coal matrix swelling, as shown from Equation (5-12).
0.0
0.5
1.0
1.5
2.0
2.5
0 2 4 6 8 10 12
Vo
lum
etr
ic s
we
llin
g s
tra
in (
%)
Pore pressure (MPa)
Experimental Data
Experimental Data
Experimental Data
Modeled Data
Modeled Data
Modeled Data
(CO2)
(CH4)
(N2)
(CO2)
(CH4)
(N2)
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-11
It was considered that the increase in effective stress is responsible for the permeability
change.
(a) (b)
Figure 5-4. Illustration of contact area change due to gas sorption: (a) pre-swelling stress
state; (b) post-swelling stress state. bA and
fA are the effective contact areas before and
after swelling, respectively.
5.2.3 Development of coal permeability model
Combining the effective stress effect (Equation (5-10)) with the swelling strain effect
(Equation (5-11)) gives the resultant fracture aperture
s
f
ef
ee fC
b
bb
b
bbb
00
,00
0
,0
0 )exp( (5-13)
where 0f is the initial fracture porosity, defined as sbf /3 00 .
For a negligibly small residual fracture aperture (0,0 bb e and 0,0 eb ), Equation (5-13)
can be simplified to
s
f
ef
fCbb
0
0 )exp( (5-14)
Simplifying the above equation yields
sfef SCbb )exp(0 (5-15)
where 0/ ff fS .
Based on the cubic law between aperture change and permeability (Mckee et al., 1988;
Seidle and Huitt, 1995), the permeability change can be expressed as
p
σ
p p p p p p p
σ
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-12
33
00
)exp( sfef SCb
b
k
k
(5-16)
Coal permeability model as shown in Equation (5-16) can be extended to the three-
dimensional case. In the analysis of coal permeability the fractured coal mass is treated as a
discontinuous medium comprising both matrix and fractures (cleats). The individual matrix
blocks are represented by cubes and may behave isotropically with regard to
swelling/shrinkage, and mechanical deformability (Liu et al., 1999). The cleats are the three
orthogonal fracture sets and may also have different apertures and mechanical properties
ascribed to the different directions. Changes in coal permeability are determined by the
redistribution of effective stresses or strains due to changed conditions such as gas
injection. Typically, stresses and strains evolve at different rates in the different Cartesian
directions, and result in anisotropic permeabilities. In simulation study, the following 3D
permeability model can be implemented into numerical models
33
00
)exp( sxfexf
xx SCb
b
k
k
(5-17a)
33
00
)exp( syfeyf
yySC
b
b
k
k
(5-17b)
33
00
)exp( szfezf
zz SCb
b
k
k
(5-17c)
where ib and
ei are the cleat opening and effective stress in i direction, respectively.
zyxi ,, .
5.2.4 Physical meaning of sensitivity ratio
The initial fracture porosity, 0f , represents the fractured extent of the coal media, and the
partition factor, f , defines the influence of both injection gas components and boundary
conditions. Different gas components with different boundary conditions may have
different partition magnitudes.
In order to explain the physical meaning of the new parameter,fS , the fracture strain
change,bs , induced by coal swelling only is defined as
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-13
sf
swelling
bs Sb
b
0
(5-18)
This can be simplified into
0ssf
bs
s
bsfS
(5-19)
where sf is the volumetric free swelling strain at the final stage, and
0s is the volumetric
free swelling strain at the initial stage.
Therefore, fS represents the ratio of fracture strain change to the incremental volumetric
swelling strain, defined as a sensitivity ratio in this study. If the boundary conditions are the
same, a largerfS value means that cleat aperture change is more sensitive to sorption-
induced strain. In our experimental tests, fS varies from 6.82 to 54.8.
5.3 Permeability model evaluation
A series of gas flow-through experiments have been carried out all under constant pressure
difference conditions (Chen et al., 2011), which were defined as the difference between
confining stress and pore pressure. First, the effective stress coefficient is measured for the
non-adsorbing gas (helium) flow-through experiments. In these experiments, the impact of
gas sorption is null and any permeability alteration is considered to be due to the variation
in the effective stress coefficient. Second, the change in permeability resulting from the
non-adsorbing gas experiments is used to calibrate the subsequent experiments using
adsorbing gasses (CO2 and CH4) where the effect of sorption-induced strain alone, on
permeability change, is obtained. The measured two sets of corrected data (core No.01 and
No.02) and another two sets of experimental data from core Anderson 01 (Robertson and
Christiansen, 2007) and core Sulcis Coal (Pini et al., 2009) were used to evaluate the newly
developed permeability model in this work. Values of the volumetric swelling parameter as
listed in Table 5-1 were taken directly from these references. These values were obtained
through matching experimental data with the Langmuir curve type of strain versus pore
pressure (Robertson and Christiansen, 2007).
Table 5-1. Parameter values obtained from experimental data matching
Parameter Core No.01 Core No.02 Anderson 01 Sulcis coal
Langmuir strain constant 0.052 0.045 0.0353 0.049
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-14
for CO2 (%)
Langmuir pressure constant
for CO2 (MPa) 5.20 3.20 3.83 7.25
Langmuir strain constant
for CH4 (%) 0.030 - 0.0168 -
Langmuir pressure constant
for CH4 (MPa) 2.96 - 6.11 -
5.3.1 Permeability model verification
Model results were compared with experimental data for cores No.01 and 02. Effects of
sorption-induced strain alone on permeability change were investigated in this comparison.
In this comparison, only the sensitivity ratio, fS , is adjustable in this match, and matching
results are shown in Figures 5-5 and 5-6.
Both data matches verify the validity of this newly developed coal permeability model. It
can be seen that the sensitivity ratio, fS , increases with the effective stress. This
observation indicates that under a higher effective stress, larger fracture opening change is
induced, and a small change of fracture aperture could cause a dramatic change in the
0bbswelling
ratio.
(a)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 2 4 6 8 10 12 14
Pe
rme
ab
ilit
y (
mD
)
Pore pressure (MPa)
Experimental Data
Experimental Data
Experimental Data
Modeled Data
Modeled Data
Modeled Data
(σe=2.0 MPa)
(σe=4.0 MPa)
(σe=6.0 MPa)
(σe=2.0 MPa)
(σe=4.0 MPa)
(σe=6.0 MPa)
Sf =6.82
Sf =7.21
Sf =6.90
Flow Fluid: CH4
Core: No.01
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-15
(b)
Figure 5-5. Comparisons of experimental permeability with the model results for core
No.01: (a) adsorbing gas CH4; (b) adsorbing gas CO2.
Figure 5-6. Comparison of experimental permeability with the modeled results for core
No.02.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 2 4 6 8 10 12 14
Pe
rme
ab
ilit
y (
mD
)
Pore pressure (MPa)
Experimental Data
Experimental Data
Experimental Data
Modeled Data
Modeled Data
Modeled Data
Flow Fluid: CO2
Core: No.01(σe=2.0 MPa)
(σe=4.0 MPa)
(σe=6.0 MPa)
(σe=2.0 MPa)
(σe=4.0 MPa)
(σe=6.0 MPa)
Sf =7.25
Sf =8.70
Sf =9.93
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1 2 3 4 5 6
Pe
rme
ab
ilit
y (
mD
)
Pore pressure (MPa)
Experimental Data
Experimental Data
Experimental Data
Modeled Data
Modeled Data
Modeled Data
Flow Fluid: CO2
Core: No.02
(σe=2.0 MPa)
(σe=3.0 MPa)
(σe=4.0 MPa)
(σe=2.0 MPa)
(σe=3.0 MPa)
(σe=4.0 MPa)
Sf =11.32
Sf =15.44
Sf =17.03
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-16
5.3.2 Comparison with other permeability models
In this section, experimental data from core Anderson 01 (Robertson and Christiansen,
2007) and core Sulcis Coal (Pini et al., 2009) were used to compare this developed
permeability model with other widely used permeability models, including updated Palmer-
Mansoori (P-M) model, Shi-Durucan (S-D) model, Cui-Bustin (C-B) model (Shi and
Durucan, 2004; Cui and Bustin, 2005; Palmer et al., 2007). For coal Anderson 01,
experimental data for CO2 and CH4 were chosen. The confining pressure was 6.895 MPa
(1,000 psi) for all experiments, and injection pressure varied from 0.5 MPa to 5.6 MPa. For
Sulcis Coal, the confining stress was 10.0 MPa and the injection pressure increased from
0.49 to 7.75 MPa. These comparisons are to benchmark the performance of our model
against others.
Coal swelling parameters from laboratorial tests were used, as listed in Table 5-1. Because
the cleat compressibility was not given in both references, both the fracture compressibility
and the sensitivity ratio, fS , were considered as the variables for our permeability model.
For other permeability models, the physical properties of Young's modulus and Poisson's
ratio are recovered directly from the experiments (Robertson and Christiansen, 2007; Pini
et al., 2009), and fracture compressibility was considered to be variable for both S-D model
and C-B model. Fracture porosity is the matching parameter for P-M model. The best
matching parameters for each model are listed in Table 5-2, and the comparison results are
shown in Figures 5-7 and 5-8.
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-17
(a)
(b)
Figure 5-7. Comparison of experimental permeability data (Robertson and Christiansen,
2007) with the modeled ones for core Anderson 01: (a) adsorbing gas CO2; (b) adsorbing
gas CH4.
0.5
0.6
0.7
0.8
0.9
1.0
1.1
1.2
0 1 2 3 4 5 6
Pe
rme
ab
ilit
y r
ati
o,
k/k
0
Pore pressure (MPa)
Experimental Data
New Model
S-D Model
P-M model
C-B model
Flowing fluid: CO2
Core: Anderson 01
0.5
0.6
0.7
0.8
0.9
1.0
1.1
1.2
0 1 2 3 4 5 6
Pe
rme
ab
ilit
y r
ati
o,
k/k
0
Pore pressure (MPa)
Experimental Data
New Model
S-D Model
P-M model
C-B model
Flowing fluid: CH4
Core: Anderson 01
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-18
Figure 5-8. Comparison of experimental permeability data (Pini et al., 2009) with the
modeled ones for core Sulcis.
Table 5-2. Matching parameters used in the comparison of permeability models
Sample
Model name
Physical
property S-D model Our model P-M model C-B model
Anderson
01 (CO2)
ν= 0.3
E=1.38 GPa Cf=0.0142 MPa-1
Sf =35.66
βCf=0.0893 MPa-1
g=1.0
Φf0=3.18% Cf=0.029 MPa-1
Anderson
01 (CH4)
ν= 0.3
E=1.38 GPa Cf=0.0101 MPa-1
Sf =54.78
βCf=0.065 MPa-1
g=1.0
Φf0=3.55% Cf=0.165 MPa-1
Sulcis
coal
ν= 0.26
E=1.12 GPa Cf=0.01 MPa-1
Sf =36.78
βCf=7.63 MPa-1
g=1.0
Φf0=1.0% Cf=0.001MPa-1
For the compared permeability models, as shown in Figures 5-7 and 5-8, the total swelling
strain is used to calculate permeability variation. The experimental data show that sorption-
induced strain only plays a dominant role at low pressures (permeability reduction), as
shown in Figures 5-7 and 5-8, and the pore pressure induced effective stress change takes
over the dominant role (permeability increase) at higher pore pressures. Therefore, these
models are not capable of replicating this apparently anomalous behaviour if the total
0.0
1.0
2.0
3.0
4.0
5.0
6.0
0 1 2 3 4 5 6 7 8
Pe
rme
ab
ilit
y r
ati
o,
k/k
0
Pore pressure (MPa)
Experimental Data
New Model
S-D Model
P-M model
C-B Model
Flowing fluid: CO2
Core: Sulcis Coal
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-19
swelling strain data are adopted since the uniaxial-strain assumption is used in these models
while the experimental conditions are not uniaxial strain.
For our model, we considered that only part of the total swelling strain contributes to the
cleat aperture change, while the remaining part contributes to coal bulk deformation. The
effect of swelling strain on permeability change is evaluated by a partition factor, as defined
in Equation (5-11). We believe that this assumption adequately reflects the mechanism for
the interaction between coal swelling strain and permeability change, and that is why this
developed model is capable of replicating this behaviour.
In order to better explain this model, the relationship between cleat porosity and the
partition ratio of total swelling strain contributing to cleat aperture change was listed and
plotted in Table 5-3 and Figure 5-9, respectively. The data were based on the fitting results
from Figures 5-5, 5-7 and 5-8, and this comparison is to show what percentage of total
swelling strain contributes to the permeability change. Although the cleat porosity term
does not directly appear in the permeability model, it is included in the fS term,
0/ ff fS , as defined in Equation (5-15).
Table 5-3. Data for porosity and partition ratio
Porosity (%)
Partition ratio ( f )
Core No.01 (CO2) Anderson 01 (CH4) Sulcis Coal
1.7 0.169 0.931 0.625
1.5 0.149 0.822 0.552
1.3 0.129 0.712 0.478
1.1 0.109 0.603 0.405
0.9 0.089 0.493 0.331
0.7 0.070 0.383 0.257
0.5 0.050 0.274 0.184
0.3 0.030 0.164 0.110
0.1 0.010 0.055 0.037
0.08 0.008 0.044 0.029
0.06 0.006 0.033 0.022
0.04 0.004 0.022 0.015
0.02 0.002 0.011 0.007
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-20
Figure 5-9. Partition ratio of total sorption-induced strain vs. cleat porosity.
As demonstrated in Figure 5-9, the partition ratio is linearly related to cleat porosity change.
Larger cleat porosity value is accompanied by a higher partition ratio, which means more
total swelling strain is absorbed by the cleat aperture system. For instance, when the cleat
porosity is 1.0%, there is 54.78% of swelling strain contributing to cleat aperture change,
but this partition ratio decreases to 5.478% when the cleat porosity reduces to 0.1%.
All three sets of matches have illustrated that using the total swelling strain to calculate the
permeability change could dramatically overestimate its contribution, which clearly
demonstrates the contribution of this work.
5.4 Model implementation
In our previous studies (Zhang et al., 2008; Chen et al., 2009 and 2010; Liu et al., 2010a,
2010b; Wu et al., 2010 and 2011), a series of single poroelastic, equivalent poroelastic, and
dual poroelastic models were developed to simulate the interactions of multiple processes
triggered by the injection or production of both single gas and binary gas. Many studies
have also been carried out by other researchers (Cui et al., 2007; Bustin et al., 2008). In
order to reproduce the typical enigmatic behaviours of coal permeability evolution with gas
injection, we applied the new developed permeability model to a coupled 3D finite element
0.001
0.01
0.1
1
0.01 0.1 1 10
Pa
rtit
ion
ra
tio
, f
(-)
Cleat porosity (%)
Anderson 01
Sulcis Coal
Core No.01
Sf= 54.78
Sf= 36.78
Sf= 9.93
(CH4)
(CO2)
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-21
numerical model to simulate the performance of CH4 injection under stress-controlled
conditions.
5.4.1 Model descriptions
This numerical model fully couples coal geomechanical deformation, gas flow, and gas
adsorption/desorption induced coal matrix swelling/shrinkage processes (see Liu et al.
(2010b) for details). The core size is 45.0 mm in diameter and 105.5 mm in length with CH4
injection at the left-hand side. Coal is initially saturated with CH4 with pressure of 0.5MPa.
A constant injection pressure boundary condition is specified from the left side with the
value of 7.0MPa, as shown in Figure 5-10. Input parameters for this simulation are listed in
Table 5-4.
This example is to investigate the sensitivity of transient permeability with CH4 injection to
different coal physical properties as well as sorption parameters, and a series of injection
conditions was simulated as listed in Table 5-3. Simulation results were presented in terms
of (1) impacts of confining stress, (2) impacts of swelling strain, (3) impacts of fracture
compressibility, (4) impacts of effective stress coefficient, and (5) impacts of sensitivity
factor. A reference point with the coordinate (80mm, 0, 0) started from the injection side
was chosen to study the evolution of coal permeability and pore pressure in terms of
different coal parameters. Simulation results were presented in Figures 5-11 through 5-15.
Figure 5-10. Numerical simulation model under controlled stress conditions. axial and
radial represent the applied stress in axial and redial directions, respectively. The symbols
on the right hand side of the figure represent the constrained deformation in horizontal
direction, but free to move in other directions.
Sample length=105.5 mm
Sample diameter=45.0mm
radial
radial
CH4
Injection axial
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-22
Table 5-4. Properties used in the simulations
Parameter Value
Coal density, kg/m3 1250
Coal Young's modulus, GPa 0.791
Poisson's ratio 0.418
Effective stress coefficient 0.945
Fracture compressibility, MPa-1 0.0669
Methane viscosity, µPa-1 11.554
Initial gas pressure, MPa 0.5
Maximum volumetric swelling strain 0.03
Maximum adsorption gas volume, m3/ton 27.0
Langmuir pressure constant, MPa 2.96
Coal matrix porosity, % 5.0
Coal fracture porosity, % 0.5
Initial permeability, md 1.0
Sensitivity factor (fS ) 30.0
Table 5-5. Sensitivity investigation of permeability and pore pressure responses to CH4
injection under different conditions
Scenario Parameter
Case 1: Impacts of hydrostatic confining stress
MPa0.10
MPa0.9
MPa0.8
Case 2: Impacts of swelling strain
%0.4
%0.3
%0.2
L
L
L
Case 3: Impacts of fracture compressibility
1
1
1
MPa0869.0
MPa0669.0
MPa0469.0
f
f
f
C
C
C
Case 4: Impacts of effective stress coefficient
0.1
914.0
8.0
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-23
Case 5: Impacts of sensitivity factor
45
30
15
f
f
f
S
S
S
5.4.2 Simulation results and analysis
Impact of confining stress: The impact of confining stress on the evolution of coal
permeability is shown in Figure 5-11. Different confining stresses represent different coal
seam depth, which can be obtained by multiplying the pressure gradient with the reservoir
depths. Assuming the gas pressure is applied on non-adsorbing coal medium, the effective
stress increases with increasing confining stress. Therefore, more reduction in coal
permeability is achieved when the confining stress is higher. However, this initial effect has
been eliminated when we plot the permeability ratio starting from 1.0 for all three cases.
Assuming the same gas pressure condition is applied to an adsorbing coal, the sensitivities
of coal permeability are regulated by the initial effective stress when coal swelling
parameters are maintained unchanged. Therefore, when the confining stress is equal to
10MPa, the reduction in coal permeability is more significant, as shown in Figure 5-11.
Figure 5-11. Evolution of coal permeability and pore pressure under different magnitudes
of confining stresses.
0
1
2
3
4
5
6
7
8
0
0.2
0.4
0.6
0.8
1
1.2
0 5 10 15 20 25 30 35 40 45
Po
re p
res
su
re (
MP
a)
Pe
rme
ab
ilit
y r
ati
o,
k/k
0
Time (hour)
(σ=8.0 MPa)
(σ=9.0 MPa)
(σ=10.0 MPa)
Pressure(σ=8.0 MPa)
Pressure (σ=9.0 MPa)
Pressure (σ=10.0 MPa)
K/K0
K/K0
K/K0
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-24
Impact of swelling capacity: Studies from Robertson (2005) have shown that there is
large difference between the strains induced by adsorption of different gasses in coals, as
illustrated in Figure 5-3. Different maximum volumetric swelling strains used in this study
represent different gas adsorbed into coal. The influence of coal maximum swelling strain
on the evolution of both permeability and pore pressure were shown in Figure 5-12. In all
of these simulations, the initial effective stress is same for all cases. Therefore, we can
assume that when gas pressure is applied in an adsorbing coal, the sensitivities of coal
permeability are regulated by the maximum swelling strain only. Model results suggest that
for coal seam with a larger swelling capacity, the reduction in coal permeability is much
more significant, and this reduction in turn affects the pore pressure evolution.
Figure 5-12. Evolution of coal permeability and pore pressure under different magnitudes
of the maximum volumetric swelling strain constant.
Impact of fracture compressibility: The impact of compressibility is shown in Figure 5-
13. When the confining stress is kept as constant, both the effective stress effect and the
swelling strain effect are defined as a function of gas pressure. In this set of simulations, the
contribution of the effective stress to the enhancement in coal permeability is determined
by coal compressibility because the swelling factor, fS , is kept unchanged. When
fC is
higher, the permeability enhancement takes over the permeability reduction. Model results
are consistent with these conceptual analyses.
0
1
2
3
4
5
6
7
8
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
0 5 10 15 20 25 30 35 40 45
Po
re p
res
su
re
(MP
a)
Pe
rme
ab
ilit
y r
ati
o,
k/k
0
Time (hour)
Pressure
Pressure
Pressure
K/K0 (εL=2.0%)
K/K0 (εL=3.0%)
K/K0 (εL=4.0%)
(εL=2.0%)
(εL=3.0%)
(εL=4.0%)
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-25
Figure 5-13. Evolution of coal permeability and pore pressure under different magnitudes
of fracture compressibility.
Impact of effective stress coefficient: Figure 5-14 shows the profiles of permeability
evolution with different effective stress coefficients. Based on the effective stress principle,
effective stress increases as the effective stress coefficient decreases. Therefore, a smaller
effective stress coefficient will result in a larger reduction in the coal permeability
When the gas pressure is applied in an adsorbing and swelling coal, the sensitivities of coal
permeability are regulated by the initial effective stress coefficient when coal swelling
parameters are maintained unchanged. Therefore, when the effective stress coefficient is
equal to 0.8, more reduction in coal permeability is observed, as demonstrated in Figure 5-
14.
0
1
2
3
4
5
6
7
8
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
0 5 10 15 20 25 30 35 40 45
Po
re p
res
su
re
(MP
a)
Pe
rme
ab
ilit
y r
ati
o,
k/k
0
Time (hour)
Pressure
Pressure
Pressure
K/K0 (Cf=0.0469MPa-1)
K/K0 (Cf=0.0669MPa-1)
K/K0 (Cf=0.0869MPa-1)
(Cf=0.0469MPa-1)
(Cf=0.0669MPa-1)
(Cf=0.0869MPa-1)
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-26
Figure 5-14. Evolution of coal permeability and pore pressure under different magnitudes
of coal effective stress coefficient.
Impact of sensitivity factor: The impact of sensitivity factor is shown in Figure 5-15. As
stated before, fS represents the ratio of fracture aperture strain to swelling strain
incremental. When the confining stress is kept constant, both the effective stress effect and
the swelling strain effect are defined as a function of gas pressure. In this example, the
contribution of the swelling strain to the reduction in coal permeability is determined by
the sensitivity factor, fS . When
fS is higher, the permeability reduces more. Model results
are consistent with these conceptual analyses.
0
1
2
3
4
5
6
7
8
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
0 5 10 15 20 25 30 35 40 45
Po
re p
res
su
re (
MP
a)
Pe
rme
ab
ilit
y r
ati
o,
k/k
0
Time (hour)
Pressure
Pressure
Pressure
K/K0 (α=0.8)
K/K0 (α=0.945)
K/K0 (α=1.0)
(α=0.8)
(α=0.945)
(α=1.0)
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-27
Figure 5-15. Evolution of coal permeability and pore pressure under different magnitudes
of sensitivity factor.
5.5 Conclusions
Coal permeability models are required to define the transient characteristics of permeability
evolution in fractured coals. A broad variety of models have evolved to represent the
effects of sorption, swelling and stresses on the dynamic evolution of permeability. These
models can be classified into two groups: permeability models under conditions of uniaxial
strain such as Palmer-Mansoori (P-M), Shi-Durucan (S-D) and Cui-Bustin (C-B) models,
and permeability models under conditions of variable stress such as the one developed in
this study.
Although laboratory experiments are conducted under controlled conditions of stresses,
analyses of laboratory observations are normally conducted by using permeability models
under conditions of uniaxial strain. The inconsistency between experimental conditions and
modelling conditions is the reason why permeability models under conditions of uniaxial
strain cannot match the laboratory observations well as demonstrated in this study.
Permeability models under uniaxial strain are more appropriate for the overall behaviour of
coal gas reservoirs under typical in situ conditions while models representing variable stress
0
1
2
3
4
5
6
7
8
9
0
0.5
1
1.5
2
2.5
0 5 10 15 20 25 30 35 40 45
Po
re p
res
su
re
(MP
a)
Pe
rme
ab
ilit
y r
ati
o,
k/k
0
Time (hour)
K/K0 (Sf=15)
K/K0 (Sf=30)
K/K0 (Sf=45)
P (Sf=15)
P (Sf=30)
P (Sf=45)
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-28
conditions are more appropriate for behaviour examined under typical laboratory
conditions.
In this study, a phenomenological permeability model has been developed to explain why
coal permeability decreases even under the unconstrained conditions of variable stress.
Unlike permeability models under the uniaxial strain condition, this model under
conditions of variable stress is effective-stress based and can be used to recover the
important nonlinear responses due to the effective stress effects when mechanical
influences are rigorously coupled with the gas transport system. The consistency between
experimental conditions and modelling conditions is the reason why this model can match
the laboratory observations reasonably well.
Our modelling results illustrate that coal permeability profiles under the controlled stress
conditions are regulated by the following five factors: (1) confining stress. When coal
swelling parameters remain unchanged, coal permeability profiles are regulated by the
initial effective stress. Coal permeability reduces initially, recovers and then reaches 534 s
the final equilibrium magnitude. When the confining stress is higher, the final equilibrium
coal permeability is much lower than the initial permeability; (2) swelling capacity. When
confining stress conditions remain unchanged, coal permeability profiles are regulated by
coal swelling capacity. Coal permeability reduces initially, recovers and then reaches the
final equilibrium magnitude. When the swelling capacity is higher, the final equilibrium coal
permeability is much lower than the initial permeability; (3) fracture compressibility. When
the confining stress is kept as constant, both the effective stress effect and the swelling
strain effect are defined as a function of gas pressure. Under these conditions, when the
facture compressibility is higher, the permeability enhancement due to the decrease in
effective stress may take over the permeability reduction due to swelling; (4) effective stress
Coefficient. The reduction in coal permeability is larger when the effective stress coefficient
is lower because the effective stress increases as the effective stress coefficient decreases;
and (5) sensitivity factor. The sensitivity factor represents the ratio of fracture aperture
strain to swelling strain incremental. When the sensitivity factor is higher, the reduction in
coal permeability is more significant.
This study demonstrated the crucial role of the consistency between experimental
conditions and modelling conditions and the rigorous coupling between coal mechanical
deformation and gas transport in the evaluation of coal permeability observations.
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-29
5.6 Acknowledgments
This work was supported by WA:ERA, the Western Australia CSIRO-University
Postgraduate Research Scholarship, National Research Flagship Energy Transformed Top-
up Scholarship, and by NIOSH under contract 200-2008-25702. These supports are
gratefully acknowledged.
5.7 References
Berryman, J.G., 2006. Estimates and rigorous bounds on pore-fluid enhanced shear modulus in poroelastic
media with hard and soft anisotropy. International Journal of Damage Mechanics 15, 133-167.
Bustin, R.M., Cui, X., Chikatamarla, L., 2008. Impacts of volumetric strain on CO2 sequestration in coals and
enhanced CH4 recovery. AAPG Bulletin 92, 15-29.
Chen, Z., Liu, J., Connell, L., Pan, Z., Zhou, L., 2008. Impact of effective stress and CH4-CO2 counter-
diffusion on CO2 enhanced coalbed methane recovery, SPE Asia Pacific Oil and Gas Conference and
Exhibition, Perth, Australia.
Chen, Z., Liu, J., Elsworth, D., Connell, L., Pan, Z., 2009. Investigation of CO2 injection induced coal-gas
interactions, 43rd U.S. Rock Mechanics Symposium & 4th U.S. - Canada Rock Mechanics Symposium.
American Rock Mechanics Association, Asheville, North Carolina.
Chen, Z., Liu, J., Elsworth, D., Connell, L.D., Pan, Z., 2010. Impact of CO2 injection and differential
deformation on CO2 injectivity under in-situ stress conditions. International Journal of Coal Geology
81, 97-108.
Chen, Z., Pan, Z., Liu, J., Connell, D.L., Elsworth, D., 2011. Effect of the effective stress coefficient and
sorption-Induced strain on the evolution of coal permeability: Experimental observations.
International Journal of Greenhouse Gas Control 5, 1284-1293.
Clarkson, C.R., 2008. Case study: Production data and pressure transient analysis of Horseshoe Canyon CBM
Wells, CIPC/SPE Gas Technology Symposium 2008 Joint Conference. Society of Petroleum
Engineers, Calgary, Alberta, Canada.
Connell, L.D., 2009. Coupled flow and geomechanical processes during gas production from coal seams.
International Journal of Coal Geology 79, 18-28.
Connell, L.D., Lu, M., Pan, Z., 2010. An analytical coal permeability model for tri-axial strain and stress
conditions. International Journal of Coal Geology 84, 103-114.
Cui, X., Bustin, R.M., Chikatamarla, L., 2007. Adsorption-induced coal swelling and stress: Implications for
methane production and acid gas sequestration into coal seams. J. Geophys. Res. 112, B10202.
Cui, X., Bustin, R.M., 2005. Volumetric strain associated with methane desorption and its impact on coalbed
gas production from deep coal seams. AAPG Bulletin 89, 1181-1202.
Daley, T.M., Schoenberg, M.A., Rutqvist, J., Nihei, K.T., 2006. Fractured reservoirs: An analysis of coupled
elastodynamic and permeability changes from pore-pressure variation. Geophysics 71, O33-O41.
Freed, A.D., 1995. Natural strain. J. Eng. Mater. Technol. 117, 379-385.
Gilman, A., Beckie, R., 2000. Flow of coal-bed methane to a gallery. Transport in Porous Media 41, 1-16.
Gray, I., 1987. Reservoir engineering in coal seams: Part 1-The physical process of gas storage and movement
in coal seams. SPE Reservoir Engineering 2, 28-34.
Harpalani, S., Chen, G., 1997. Influence of gas production induced volumetric strain on permeability of coal.
Geotechnical and Geological Engineering 15, 303-325.
Harpalani, S., Schraufnagel, R.A., 1990. Shrinkage of coal matrix with release of gas and its impact on
permeability of coal. Fuel 69, 551-556.
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-30
Izadi, G., Wang, S., Elsworth, D., Liu, J., Wu, Y., Pone, D., 2011. Permeability evolution of fluid-infiltrated
coal containing discrete fractures. International Journal of Coal Geology 85, 202-211.
Jaeger, J.C., Cook, N.G.W., Zimmerman, R.W., 2007. Fundamentals of rock mechanics, 4 ed. Oxford:
Blackwell.
Liu, H.-H., Rutqvist, J., 2010. A new coal-permeability model: Internal swelling stress and fracture–matrix
interaction. Transport in Porous Media 82, 157-171.
Liu, H.-H., Rutqvist, J., Berryman, J.G., 2009. On the relationship between stress and elastic strain for porous
and fractured rock. International Journal of Rock Mechanics and Mining Sciences 46, 289-296.
Liu, J., Elsworth, D., Brady, B.H., 1999. Linking stress-dependent effective porosity and hydraulic
conductivity fields to RMR. International Journal of Rock Mechanics and Mining Sciences 36, 581-596.
Liu, J., Chen, Z., Elsworth, D., Miao, X., Mao, X., 2010a. Evaluation of stress-controlled coal swelling
processes. International Journal of Coal Geology 83, 446-455.
Liu, J., Chen, Z., Elsworth, D., Miao, X., Mao, X., 2010b. Linking gas-sorption induced changes in coal
permeability to directional strains through a modulus reduction ratio. International Journal of Coal
Geology 83, 21-30.
Liu, J., Chen, Z., Elsworth, D., Qu, H., Chen, D., 2011. Interactions of multiple processes during CBM
extraction: a critical review. International Journal of Coal Geology 87, 175-189.
Ma, Q., Harpalani, S., Liu, S., 2011. A simplified permeability model for coalbed methane reservoirs based on
matchstick strain and constant volume theory. International Journal of Coal Geology 85, 43-48.
Massarotto, P., Golding, S.D. and Rudolph, V., 2009. Constant volume CBM reservoirs: an important
principle, International Coalbed Methane Symposium, Tuscaloosa, Alabama.
Mavko, G., Jizba, D., 1991. Estimating grain-scale fluid effects on velocity dispersion in rocks. Geophysics 56,
1940-1949.
McKee, C.R., Bumb, A.C., Koenig, R.A., 1988. Stress-dependent permeability and porosity of coal and other
geologic formations. SPE Formation Evaluation 3, 81-91.
Palmer, I., Mansoori, J., 1996. How permeability depends on stress and pore pressure in coalbeds: A new
model, SPE Annual Technical Conference and Exhibition. Society of Petroleum Engineers, Inc.,
Denver, Colorado.
Palmer, I.D., Mavor, M., Gunter, B., 2007. Permeability changes in coal seams during production and
injection, Intl. Coalbed Methane Symposium, University of Alabama, Tuscaloosa, Alabama. Paper
0713.
Pan, Z., Connell, L.D., 2007. A theoretical model for gas adsorption-induced coal swelling. International
Journal of Coal Geology 69, 243-252.
Pan, Z., Connell, L.D., 2011a. Modelling of anisotropic coal swelling and its impact on permeability
behaviour for primary and enhanced coalbed methane recovery. International Journal of Coal
Geology 85, 257-267.
Pan, Z., Chen, Z., Connell, L.D., Lupton, N., 2011b. Laboratory characterisation of fluid flow in coal for
different gases at different temperatures, Asia Pacific Coalbed Methane Symposium, Brisbane,
Australia.
Pekot, L.J., Reeves, S.R., 2002. Modeling the effects of matrix shrinkage and differential swelling on coalbed
methane recovery and carbon sequestration. U.S. Department of Energy, DE-FC26-00NT40924.
Pini, R., Ottiger, S., Burlini, L., Storti, G., Mazzotti, M., 2009. Role of adsorption and swelling on the
dynamics of gas injection in coal. J. Geophys. Res. 114, B04203.
Robertson, E.P., 2005. Modeling permeability in coal using sorption-induced strain data, SPE Annual
Technical Conference and Exhibition. Society of Petroleum Engineers, Dallas, Texas.
Robertson, E.P., Christiansen, R.L., 2006. A permeability model for coal and other fractured, sorptive-elastic
media, SPE Eastern Regional Meeting. Society of Petroleum Engineers, Canton, Ohio, USA.
School of Mechanical and Chemical Engineering
Chapter 5 The University of Western Australia
5-31
Robertson, E.P., Christiansen, R.L., 2007. Modeling laboratory permeability in coal using sorption-induced
strain data. SPE Reservoir Evaluation & Engineering 10, pp. 260-269.
Sawyer, W.K., Paul, G.W., Schraufnagel, R.A., 1990. Development and application of a 3D coalbed simulator,
International Technical Meeting Hosted Jointly by the Petroleum Society of CIM and the Society of
Petroleum Engineers., Calgary, Alberta, Canada. CIM/SPE 90-1119.
Seidle, J.R., Huitt, L.G., 1995. Experimental measurement of coal matrix shrinkage due to gas desorption and
implications for cleat permeability increases, International Meeting on Petroleum Engineering. Society
of Petroleum Engineers, Inc., Beijing, China.
Shi, J.Q., Durucan, S., 2004. Drawdown induced changes in permeability of coalbeds: A new interpretation of
the reservoir response to primary recovery. Transport in Porous Media 56, 1-16.
Shi, J.-Q., Durucan, S., 2010. Exponential growth in San Juan Basin Fruitland coalbed permeability with
reservoir drawdown: Model match and new insights. SPE Reservoir Evaluation & Engineering 13,
914-925.
Somerton, W.H., Söylemezoglu, I.M., Dudley, R.C., 1975. Effect of stress on permeability of coal.
International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts 12, 129-145.
Tonnsen, R.R., Miskimins, J.L., 2010. Simulation of deep coalbed methane permeability and production
assuming variable pore volume compressibility, Canadian Unconventional Resources and
International Petroleum Conference, Calgary, Alberta, Canada.
White, C.M., Smith, D.H., Jones, K.L., Goodman, A.L., Jikich, S.A., LaCount, R.B., DuBose, S.B., Ozdemir,
E., Morsi, B.I., Schroeder, K.T., 2005. Sequestration of carbon dioxide in coal with enhanced coalbed
methane recovery- A review. Energy & Fuels 19, 659-724.
Wu, Y., Liu, J., Chen, Z., Elsworth, D., Pone, D., 2011. A dual poroelastic model for CO2-enhanced coalbed
methane recovery. International Journal of Coal Geology 86, 177-189.
Wu, Y., Liu, J., Elsworth, D., Chen, Z., Connell, L., Pan, Z., 2010. Dual poroelastic response of a coal seam
to CO2 injection. International Journal of Greenhouse Gas Control 4, 668-678.
Zhang, H., Liu, J., Elsworth, D., 2008. How sorption-induced matrix deformation affects gas flow in coal
seams: A new FE model. International Journal of Rock Mechanics and Mining Sciences 45, 1226-1236.
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-1
Chapter 6
Roles of Coal Heterogeneity on Evolution of Coal Permeability under Unconstrained Boundary Conditions
Zhongwei Chena, Jishan Liua*, Derek Elsworthb, Jianguo Wanga, Zhejun Panc, Luke D.
Connellc
a School of Mechanical and Chemical Engineering, The University of Western Australia, WA 6009, Australia
b Department of Energy and Mineral Engineering, Penn State University, PA 16802-5000, USA
c CSIRO Earth Science and Resource Engineering, Private Bag 10, Clayton South, Victoria 3169, Australia
Abstract: Coal permeability models based constrained conditions such as constant volume
theory can successfully describe unconstrained experimental data and field observations.
However, these models have a boundary mismatch because model boundary is constrained
while experiment boundary is free displacement or unconstrained. What the mechanism is
to require such a boundary mismatch has not been well understood so far. In this study, a
full coupling approach was developed to explicitly model the interactions of coal matrix
and fracture. In this model, a matrix-fracture model is numerically investigated after
incorporating heterogeneous distributions of Young's modulus, Langmuir strain constant
in the vincity of the fracture. The impact of these local heterogeneities of coal mechanical
and swelling properties on the permeability evolution is explored. The transient
permeability evolution during gas swelling process was investigated and the difference
between the final equilibrium permeability and transient permeability was compared. It is
found that a net reduction of coal permeability is achieved from the initial no-swelling state
to the final equilibrium state. This net reduction of coal permeability increases with the
fracture (injection) pressure and is in good agreement with laboratorial data under the
unconstrained swelling conditions. Therefore, the local heterogeneity of coal in vincity of
fracture should be the mechanism of the above mismatch.
Keywords: coal permeability; gas sorption; coal swelling; coal-gas interaction; local
heterogeneity.
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-2
6.1 Introduction
The permeability of coal is a key attribute in determining coal seam methane production
and CO2 storage in coal seam reservoirs. In coal the permeability is often determined by
regular sets of fractures called cleats, with the aperture of the cleats being a key property in
the magnitude of the permeability (Connell et al., 2010). The relative roles of stress level,
gas pressure, gas composition, fracture geometry of coal, and water content are intimately
connected to the processes of gas sorption, diffusion, transport, and coal
swelling/shrinkage (Liu et al., 2011b).
Significant experimental efforts have been made to investigate coal permeability and its
evolution. Laboratory measured permeabilities of coal to adsorbing gasses, such as CH4
and CO2, are known to be lower than permeabilities to non-absorbing or lightly adsorbing
gasses such as argon and nitrogen (Durucan and Edwards, 1986; Siriwardane et al., 2009;
Somerton et al., 1975). Under constant total stress, adsorbing gas permeability decreases
with increasing pore pressure due to coal swelling (Chen et al., 2011; Mazumder and Wolf,
2008; Pan et al., 2010a; Robertson, 2005; Wang et al., 2010, 2011), and increases with
decreasing pore pressure due to matrix shrinkage (Cui and Bustin, 2005; Harpalani and
Schraufnagel, 1990; Harpalani and Chen, 1997; Seidle and Huitt, 1995). It is also
influenced by both the presence of water and the magnitude of water saturation (Han et al.,
2010, Pan et al., 2010b). One thing is in common for the above studies that they were
conducted under unconstrained boundary conditions.
A number of proposed coal permeability models have been developed to match
experimental data (Cui and Bustin, 2005; Izadi et al., 2011; Liu and Rutqvist, 2010; Liu et al.,
2011a; Palmer and Mansoori, 1996; Pekot and Reeves, 2002; Seidle and Huitt, 1995; Shi
and Durucan, 2004; Wang et al., 2009; Zhang et al., 2008). Two assumptions are applied
with the above mentioned models- uniaxial strain and constant overburden or confining
stress (Connell et al., 2010; Liu et al., 2011a). These models have been quite successful in
matching experimental data even thought the tests were conducted under stress-controlled
(unconstrained) boundary conditions. However, permeability models developed under
stress-controlled condition is incapable to match experimental data conducted under stress-
controlled conditions, particularly for the models developed with the matchstick or cubic
coal geometry. That is because matrix swelling would not affect coal permeability due to
the complete separation between matrix blocks caused by through-going fractures. In this
case, for a given fracture pore pressure, the swelling results in an increase of fracture
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-3
spacing, rather than a change in fracture aperture (Liu and Rutqvist, 2010). However, this
has not been consistent with laboratory observations that show significant effects of matrix
swelling on coal permeability under constant confining stress conditions (Chen et al., 2011;
Lin et al., 2008; Pan et al., 2010a). This behaviour remains enigmatic as the permeability of
the porous coal is determined by the effective stress only.
A few studies have been carried out on either improving current permeability models or
explaining why permeability models developed under unaixal strain condition is capable to
match experimental data conducted under stress-controlled conditions. Connell et al. (2010)
partitioned the sorption strain into bulk, pore and matrix strains in contrast to existing
approaches, and derived several different forms of the permeability models for the distinct
geometric and mechanical arrangements that can be encountered with laboratory testing.
Liu and Rutqvist (2010) believed that in reality coal matrix blocks are not completely
separated from each other by fractures but connected by the coal-matrix bridges, and
developed a new coal-permeability model for constant confining-stress conditions, which
explicitly considers fracture–matrix interaction during coal-deformation processes based on
the internal swelling stress concept. An alternative reasoning has been investigated by Liu
et al. (2010a), considering that the internal actions between coal fractures and matrix have
not been taken into consideration. Recently, Izadi et al. (2011) proposed a mechanistic
representation of coal as a collection of unconnected cracks in an elastic swelling medium,
where voids within a linear solid are surrounded by a damage zone. In the damage zone the
Langmuir swelling coefficient decreases outwards from the wall and the modulus increases
outwards from the wall. In the analysis, fluid pressures are applied uniformly throughout
the body, so it is incapable of observing the transient permeability evolution due to coal-gas
interactions during gas transport. Liu et al. (2011b) addressed the same phenomena from
different point of view, stating that coal permeability is controlled by the switching process
between local swelling and macro-swelling, and the extent of switching of coal swelling
determines coal permeability is higher or lower than initial value.
However, these studies still have three limitations can be improved: (1) they were generally
carried out on the assumption of homogeneity, where coal properties were assumed to be
same throughout the whole domain; (2) it was assumed that permeability value is only
related to pore pressure and effective stress, so with same pore pressure the permeability
value is same; (3) permeability is independent of time. These assumptions have been
conflict with many experimental observations. Maggs (1946) investigated the feature of coal
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-4
swelling, and shown that in the presence of an adsorbed film, the coal swells and a
weakening of the structure would result on adsorption. This phenomenon has also been
confirmed by Hsieh and Duda (1987). The effect of high-pressure CO2 on the
macromolecular structure of coal has been studied by Mirzaeian and Hall (2006), and
shown that the glass transition temperature of coal decreases with CO2 pressure
significantly, indicating that high-pressure CO2 diffuses through the coal matrix, causes
significant plasticization effects, and changes the macromolecular structure of coal. Similar
observation has been obtained by many other researchers (Larsen, 2004; Goodman et al.,
2005; John, 2004; Liu et al., 2010; White et al., 2005). The thermodynamics and mechanism
for this phenomenon was examined by Mirzaeian and Hall (2008). The plasticization
effects of coal adsorption have been verified by the weakening of coal mechanical strength
from experimental measurements (Ates and Barron, 1988; Ranjith et al., 2010; Viete and
Ranjith, 2006; Viete and Ranjith, 2007; Wang et al., 2011). Recently, Siriwardane et al. (2009)
found that permeability of adsorbing gas in coal is a function of gas exposure time.
The non-homogeneous feature of coal swelling has also been observed by other
approaches (Day et al., 2008; Karacan and Okandan, 2001; Karacan, 2003; Karacan, 2007)
as apparent from quantitative X-ray CT imaging and from optical methods. Gibbins et al.
(1999) examined the heterogeneity of coal samples by means of density separation and
optical and scanning electron microscopy, and found that a high degree of heterogeneity
exists between average compositions for the different density cuts within each sample,
between different particles within the same density cuts, and within the particles themselves.
Similar work has been conducted by Gathitu et al. (2009). Manovic et al. (2009) presented
the microscopic observations of coals of different rank and mineral matter content,
showing an increasing of heterogeneity with mineral matter content. Anisotropic swelling
induced by chemical heterogeneity of coal has also been observed (Douglas, 1984; French
et al., 1993; Pone et al., 2010).
As summarized above, the real behaviours of the sorption-induced swelling/shrinkage of
coal are far different from the homogeneous assumption, which is generally made for
theoretical permeability analysis. The effects of coal chemical heterogeneity and swelling
are mutual. The heterogeneity of coal brings the non-homogeneous distribution of coal
swelling strain, and meanwhile coal swelling causes the heterogeneous distribution coal
physical property (e.g. Young's modulus). In this study, it is considered that the
heterogeneities of coal physical properties and swelling strain are responsible for the
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-5
enigmatic behaviour of coal permeability reduction with adsorbing gas injection under
unconstrained conditions. To prove this, a fully coupling numerical model is conducted to
simulate the dynamic interactions between coal matrix swelling and fracture aperture
alteration, and translate these interactions to transient permeability evolution. In this
numerical model, swelling coefficient and Young's modulus are assumed to vary spatially,
and numerical predictions are then compared with observed magnitudes of permeability
change in coal. Our work is trying to explain why permeability changes with absorbing gas
injection even under stress controlled conditions.
6.2 Theoretical evaluation of coal permeability models
6.2.1 General coal permeability model
It is clear that there is a relationship between porosity, permeability and the grain-size
distribution in porous media. Chilingar (1964) defined this relationship as
2
32
)1(72
ed
k (6-1)
where k is the permeability, is porosity and ed is the effective diameter of grains. Based
on this equation, we obtain
2
0
3
00 1
1
k
k. (6-2)
When the porosity is much smaller than 1 (normally less than 10%), the second term of the
right-hand side asymptotes to unity. This yields the cubic relationship between permeability
and porosity for coal matrix
3
00
k
k (6-3)
Coal porosity can be defined as a function of the effective strain (Liu et al., 2010a, 2010b)
e
00
1 (6-4)
Substituting Equations (6-4) into (6-3) gives
3
00
1
e
k
k
(6-5)
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-6
s
s
veK
p
(6-6)
or
K
pe
(6-7)
where e is defined as the total effective volumetric strain,
v is total volumetric strain,
sKp is coal compressive strain, s is gas sorption-induced volumetric strain, and
sK
represents the bulk modulus of coal grains.
Equations (6-4) and (6-5) are models for coal porosity and permeability that are derived
based on the fundamental principles of poroelasticity. They can be applied to the evolution
of coal porosity and permeability under variable boundary conditions.
Coal porosity and permeability can be defined as a function of either effective strain (6-6)
or effective stress (6-7). However, coal porosity and permeability models may have a
variety of forms when specific conditions are imposed. Examples include:
When the change in total stress is equal to zero, 0 , both coal porosity and
permeability are independent of the total stress. Under this condition, they can be
defined as a function of gas pressure only.
Assuming coal sample is under conditions of uniaxial strain and the overburden load
remains unchanged, they can also be defined as a function of gas pressure only.
When the impact of coal fractures and gas compositions is considered, coal porosity
and permeability models can be linked to fracture parameters and gas concentrations.
6.2.2 Evaluation of permeability model under two boundary conditions
In this section, general coal permeability models are evaluated through comparing
laboratorial and in-situ measurements with theoretical solutions of the two extreme cases,
the unconstrained shrinkage/swelling model and the constrained model, as illustrated in
Figure 6-1.
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-7
(a) (b)
Figure 6-1. Schematic diagram of two extreme cases: (a) unconstrained model (free swelling
model), where constant stress conditions are applied throughout the whole process; (b)
constrained model (constant volume model), where constant volume conditions are
maintained throughout the whole process. These two cases represent the lower and upper
bounds for permeability and porosity response.
We assume that matrix blocks are completely separated from each other in coal sample.
For the unconstrained model, matrix swelling does not affect coal fracture permeability,
because for a given pore pressure coal matrix swelling results in swelling of the blocks
alone, rather than changes in fracture aperture. The ambient effective stress also exerts no
influence on matrix swelling, due to the complete separation between matrix blocks caused
by through-going fractures. However, when coal sample is completely constrained from all
directions, coal matrix swelling will be completely transferred to the reduction in fracture
apertures. For the constrained model, the entire swelling/shrinkage strain contributes to
coal permeability change provided the fractures are much more compliant than coal matrix.
Equation (6-5) is derived based on the poroelastic theory without the effect of fractures.
Therefore, the porosity should be the matrix porosity. However, when these models are
applied, the fracture porosity is actually used. Evolutions of coal permeability for both
unconstrained and constrained conditions can be defined as Equations (6-8) and (6-9),
respectively.
3
00
1
K
p
k
k
f
(6-8)
3
00
1
s
s
f K
p
k
k
(6-9)
Pre-swelling
Aperture
Post-swelling
Aperture
Pre-swelling
Aperture
Post-swelling
Aperture
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-8
Solutions of these two cases and their comparisons with typical observations are illustrated
in Figures 6-2 to 6-4. Robertson (2005) directly measured the influence of coal swelling on
permeability change with four different gasses (helium, N2, CH4 and CO2) injection. Similar
experiments have been conducted by others (Pini et al., 2009; Siriwardane et al., 2009;
Wang et al., 2010; Kiyama et al., 2011). These observations demonstrate that even under
unconstrained stress-controlled conditions the injection of adsorbing gasses reduces coal
permeability at a lower gas pressure and coal permeability may rebound at a higher gas
pressure. This observed switch in behaviour is presumably due to the dependence of coal
swelling on the gas pressure: coal swelling diminishes at high pressures. Because all of the
mentioned experimental observations were made under controlled stress conditions, they
should be equal to or close to the theoretical solution under the unconstrained swelling
condition. This has not been the case, as illustrated in Figure 6-2. These observations
indicate that although the experiments were conducted under controlled stress conditions
the experimental measurements are more closely related to those expected under constant
volume conditions. These discrepancies illustrate the obvious drawbacks of the current coal
permeability models. If a coal gas reservoir is treated as a whole, with full lateral restraint
and invariant overburden stress, its behaviour should represent components of the free
swelling/shrinkage and the constant volume models. This could explain why current coal
permeability models representing conditions of uniaxial strain condition can successfully
match some field data.
Figure 6-2. Illustration of discrepancy between model predicted coal permeability and
typical laboratory measurements for unconstrained swelling case (Liu et al., 2011a).
0k
k
110
110
p1
0p
3
00
1
K
p
k
k
f
3
00
1
s
s
f K
p
k
k
p
p
Zone of Discrepancy
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-9
The influence of coal matrix shrinkage on coal permeability has also been widely studied
and an increase with decreasing pore pressure due to matrix shrinkage has been observed
(Cui and Bustin, 2005; Harpalani and Schraufnagel, 1990; Harpalani and Chen, 1997; Seidle
and Huitt, 1995), as shown in Figure 6-3. Similar trend has obtained from field
observations. In-situ measured data show that the absolute permeability of coal gas
reservoirs increases significantly with continued gas drainage (Cherian et al., 2010; Clarkson,
2008; Sparks et al., 1995; Young et al., 1991). Comparison of field observations with both
the unconstrained and constrained models is presented in Figure 6-4. Because all of the in-
situ observations were made under unknown conditions of in-situ stress they should lie
within the bracketing behaviours. Both Figure 6-3 and Figure 6-4 demonstrate that coal gas
reservoirs behave more closely to the constrained (constant volume) case.
Figure 6-3. Illustration of discrepancy between model predicted coal permeability and
typical laboratory measurements for unconstrained shrinkage case (Liu et al., 2011a).
0k
k
110
110
p1
0p
3
00
1
K
p
k
k
f
3
00
1
s
s
f K
p
k
k
p
p
Zone of Discrepancy
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-10
Figure 6-4. Illustration of discrepancy between model predicted coal permeability and
typical in-situ observations (Liu et al., 2011a).
The analysis above has demonstrated that current coal theoretical permeability models have
so far been unsuccessful to explain the results from stress-controlled shrinkage/swelling
laboratorial tests and have only achieved some limited success in explaining and matching
in situ data. The most recent viewpoints (Izadi et al., 2011; Liu and Rutqvist, 2010) have
demonstrated that the main reason for the failure is that the impact of coal matrix–fracture
compartment interactions on the evolution of coal permeability has not been incorporated
appropriately as most of the coal permeability models are derived based on the theory of
poroelasticity.
During laboratorial tests, coal permeability is measured only when gas flow is considered
having reached equilibrium state, in which a uniform matrix swelling/shrinkage is achieved
for a homogeneous coal sample (Siriwardane et al., 2009; Wang et al., 2011). In this work,
coal permeability at this state is defined as the final equilibrium permeability. However, this
condition may never be achieved for real coal samples. A difference of the final equilibrium
(or ultimate) permeability between an ideal homogeneous coal and a real heterogeneous
coal is expected. In the following section, a simulation model is constructed to investigate
the transient permeability evolution during gas swelling process and to study the difference
between the final equilibrium permeability and transient permeability, from which the
possible reasons for permeability reduction under unconstrained conditions can be
0k
k
110
110
p1
0p
3
00
1
K
p
k
k
f
3
00
1
s
s
f K
p
k
k
p
p
Zone of Discrepancy
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-11
achieved. In this approach, the important non-linear responses of coal matrix to the
effective stress are quantified through the incorporation of heterogeneous distributions of
coal properties into complex mechanical coupling with gas transport, where swelling
coefficient and modulus vary spatially relative to the fracture void.
6.3 A heterogeneous matrix-fracture interaction model
The key to model the dynamic interactions between coal matrix swelling/shrinkage and
fracture aperture alteration is to recover important non-linear responses of coal matrix to
the effective stress. In order to achieve this goal, geomechanical influence needs to be
rigorously coupled with the gas transport process. Full coupling approach has to be chosen
to achieve this goal, from which a single set of equations (generally a large system of non-
linear coupled partial differential equations) incorporating all of the relevant physics are
solved simultaneously (Liu et al., 2011a). Full coupling is often the preferred method for
simulating multiple types of physics since it should theoretically produce the most realistic
results. Therefore, it would represent important non-linear responses due to the effective
stress effects when geomechanical influences are rigorously coupled with the gas transport
system. Over the past few years, a series of advanced modelling tools has been developed
to quantify the complex coal–gas interactions (Chen et al., 2009, Chen et al., 2010; Connell,
2009; Connell and Detournay, 2009; Gu and Chalaturnyk, 2005, 2006; Liu et al., 2010a,b,
2011b; Wu et al., 2010a,b; Zhang et al., 2008). To reproduce the typical enigmatic
behaviours of coal permeability evolution with gas injection, in the following section we
applied this full coupling approach to simulate the evolution of coal permeability under
unconstrained conditions.
6.3.1 Numerical model implementation
In this section, a simulation model was constructed to investigate the permeability change
under unconstrained conditions. It was considered that the interactions of the fractured
coal mass where cleats do not create a full separation between adjacent matrix blocks but
where solid rock bridges are present, as illustrated in Figure 6-5(a). We accommodate the
role of swelling strains both over contact bridges that hold cleat faces apart and over the
non-contacting span between these bridges. The effects of swelling act competitively over
these two components: increasing porosity and permeability due to swelling of the bridging
contacts but reducing porosity and permeability due to the swelling of the intervening free-
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-12
faces. The influence of effective stress and swelling response for a rectangular crack is
examined, and we consider a single component part removed from the array where the
appropriate boundary conditions are for uniform displacement along the boundaries. This
represents the symmetry of the displacement boundary condition mid-way between flaws
as shown in Figure 6-5(b). The change in aperture due to the combined influence of coal
sorption-induced swelling and effective stress change is calculated during gas transport
process, from which the transient permeability evolution is obtained. The cubic
relationship between permeability and aperture change was chosen to calculate permeability
ratio evolution.
The simulation model geometry is 1.0cm by 2.0cm with a fracture located at the centre of
the model. The length and opening of the fracture are 0.3cm and 0.05mm, respectively. For
coal deformation model, the right and bottom sides are constrained in the normal direction
to honor symmetry and the other two sides are stress controlled, as shown in Figure 6-5(b).
For the gas transport model (CO2 is used for this study), coal sample is initially saturated
with gas with 0.2 MPa pressure, and a constant injection pressure is specified at the
boundaries of fracture. Input parameters for this simulation are listed in Table 6-1.
(a) Multiple fracture compartment model (b) Single fracture compartment model
Figure 6-5. Numerical model for permeability change under the unconstrained boundary
condition
Table 6-1. Parameters used in the numerical model
Parameter Value
Porosity, % 5.0
Matrix permeability, m2 5.0×10-22
Viscosity, Pa*s 1.228×10-5
2.0 cm
0.3 cm
0.05 mm
1.0 cm
Co
nfi
nin
g S
tres
s
Confining Stress
Fracture
A
B
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-13
Young's modulus, GPa 3.45
Poisson's ratio 0.4
Maximum volumetric swelling strain of CO2 0.03
CO2 Langmuir sorption constant, m3/kg 0.0132
CO2 Langmuir pressure constant, MPa 3.96
Coal density, kg/m3 1500
Young's modulus softening coefficient, 0.75
Swelling strain reduction factor, 0.25
6.3.2 Performance for a homogeneous coal
Firstly coal matrix is assumed to be homogeneous, where Young's modulus and Poisson‘s
ratio are constants and the Langmuir strain constant is same throughout the domain. The
simulation scenarios are listed in Table 6-2. Typical evolution of coal permeability is shown
in Figure 6-6. As can be seen from it, coal permeability experiences a rapid reduction at the
early stage. When the modelled time is about 2000s, a switch in behaviour from
permeability reduction to recovery is observed. After this, coal permeability recovers until it
reaches the final equilibrium permeability. The final equilibrium permeability is higher than
the original one, which is inconsistent with laboratorial observations (Harpalani and
Schraufnagel, 1990; Seidle and Huitt, 1995; Levine, 1996; Robertson, 2005; Cui et al., 2007;
Karacan, 2007; Day et al., 2008; Mazumder and Wolf, 2008; Kiyama et al., 2011; Wang et
al., 2010, 2011). Permeability increase with gas injection has been obtained for all
homogenous cases as shown in Figure 6-7. Final equilibrium permeability increases with
increasing fluid pressure, Poisson‘s ratio and volume swelling strain capacity, but decreases
with rising Young‘s Modulus.
Our conceptual understanding on the modelling processes is illustrated in Figure 6-8. Prior
to the CO2 injection, the gas pressure in the fracture is equal to that in the matrix. We
define this state as the initial equilibrium, 0ppp fm . In this study, coal permeability at
this state is defined as the initial equilibrium permeability, as illustrated in Figure 6-8(a).
When CO2 is injected, the gas occupies the fracture and the gas pressure in the fracture
reaches the injection pressure almost instantly. At this stage, the maximum imbalance
between fracture pressure and matrix pressure is achieved. However, this imbalance
diminishes as the gas diffuses into coal matrix. As a consequence of the diffusion, coal
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-14
matrix swells. Initially, matrix swelling is confined in the vicinity of the fracture voids. This
localized swelling reduces the fracture aperture, and in turn reduces the fracture
permeability, as shown in Figure 6-8(b). As CO2 diffusion progresses, the swelling zone
extends further into coal matrix, and the influence of matrix swelling on the fracture
aperture weakens. As a result of the widening of the swelling zone, the fracture
permeability recovers. When the imbalance between fracture pressure and matrix pressure
diminishes completely, the final equilibrium state is achieved, as shown in Figure 6-8(c). At
the final equilibrium state, the fracture pressure is equal to the matrix pressure, i.e.,
infm ppp , where inp is the injection pressure. Coal permeability at this state is defined
as the final equilibrium permeability.
Table 6-2. List of simulation scenario for homogeneous case
Simulation
scenario
Fracture pressure
fp (MPa)
Young's
modulus 0E
Langmuir strain
constant L (%)
Poisson‘s
ratio µ(-)
Case 1 2.0, 4.0, 6.0, 8.0,
10 1.0 3.0 0.40
Case 2 10.0 0.5, 0.75,
1.0, 1.25 3.0 0.40
Case 3 10.0 1.0 1.0, 2.0,
3.0, 4.0 0.40
Case 4 10.0 1.0 3.0 0.25, 0.30,
0.35, 0.40
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-15
Figure 6-6. Numerical result of permeability evolution for the homogeneous coal.
(a)
(b)
0.2
0.4
0.6
0.8
1
1.2
1.0E+00 1.0E+01 1.0E+02 1.0E+03 1.0E+04 1.0E+05 1.0E+06
Initial EquilibriumFinal Equilibrium
Net Permeability Increase
Transitional Period
0mpfm pp
fm pyxp ,0
Confi
nin
g S
tres
s
Confining Stress
yxpm ,
Fracture
PressureInjection fp
Constant
Constant
Case sHomogeneou
L
E
110 210 310 410 510 610
Time (S)
Co
al P
erm
eab
ilit
y R
ati
o
010
Co
al p
erm
eabil
ity r
ati
o
Time (S)
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
1.0E+0 1.0E+1 1.0E+2 1.0E+3 1.0E+4 1.0E+5 1.0E+6
Dy
na
mic
pe
rme
ab
ilit
y r
ati
o (
k/k
0)
Time (s)
P=2.0MPa
P=4.0MPa
P=6.0MPa
P=8.0MPa
P=10.0MPa
Initial equilibrium
Final equilibrium
1
1.01
1.02
1.03
1.04
1.05
1.06
1.07
1.08
2 4 6 8 10
Fin
al e
qu
ilib
riu
m p
erm
ea
bilit
y r
ati
o
Pore pressure (MPa)
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
1.0E+0 1.0E+1 1.0E+2 1.0E+3 1.0E+4 1.0E+5 1.0E+6
Dy
na
mic
pe
rme
ab
ilit
y r
ati
o (
k/k
0)
Time (s)
0.50 E0
0.75 E0
E0
1.25 E0
Initial equilibrium
Final equilibrium
1
1.01
1.02
1.03
1.04
1.05
1.06
1.07
1.08
0.5 0.7 0.9 1.1 1.3
Fin
al e
qu
ilib
riu
m p
erm
ea
bilit
y r
ati
o
Young's modulus ratio (E/E0)
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-16
(c)
(d)
Figure 6-7. Evolutions of both dynamic permeability and equilibrium permeability for a
homogeneous coal under the unconstrained swelling condition. (a) Influence of different
pore pressures; (b) influence of different modulus magnitudes; (c) influence of different
Langmuir strain constants; (d) influence of different Poisson‘s ratios.
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
1.0E+0 1.0E+1 1.0E+2 1.0E+3 1.0E+4 1.0E+5 1.0E+6
Dy
na
mic
pe
rme
ab
ilit
y r
ati
o (
k/k
0)
Time (s)
Initial equilibrium
εs=1.0%
εs=2.0%
εs=3.0%
εs=4.0%
Final equilibrium
1
1.01
1.02
1.03
1.04
1.05
1.06
1.07
1.08
1 1.5 2 2.5 3 3.5 4
Fin
al e
qu
ilib
riu
m p
erm
ea
bilit
y r
ati
o
Langmuir swelling strain constant (%)
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
1.0E+0 1.0E+1 1.0E+2 1.0E+3 1.0E+4 1.0E+5 1.0E+6
Dy
na
mic
pe
rme
ab
ilit
y r
ati
o (
k/k
0)
Time (s)
Initial equilibrium
µ=0.25
µ=0.30
µ=0.35
µ=0.40
Final equilibrium
1
1.01
1.02
1.03
1.04
1.05
1.06
1.07
0.25 0.30 0.35 0.40
Fin
al e
qu
ilib
riu
m p
erm
ea
bilit
y r
ati
o
Poisson's ratio (-)
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-17
Figure 6-8. Illustration of the difference of swelling behaviours between the homogeneous
coal and the heterogeneous coal. The red dotted line represents the original location of
fracture, and blue color zone represents the gas diffusion zone.
Based on the analysis above, the final equilibrium permeability is always higher than the
initial equilibrium permeability if a uniform swelling state is achieved within coal sample.
However, laboratorial measurements show that coal equilibrium permeability is generally
much lower than the initial equilibrium permeability in low pore pressure range, may
recover but rarely exceeds the initial equilibrium permeability even at high pore pressures.
This distinct discrepancy points to that a uniform matrix swelling state has rarely been
achieved in real coal sample tests. Therefore, a difference between the ultimate
permeability for an ideal homogeneous coal and that for a real heterogeneous coal is
expected.
6.3.3 Performance for a heterogeneous coal
As summarized in the introduction part, adsorption and swelling processes have been
shown to be heterogeneous in coal (Day et al., 2008; Karacan and Okandan, 2001; Karacan,
2003; Karacan, 2007), thus characteristics for the heterogeneity may include:
Fracture Compartment
Matrix
fp fp
fm pp 0
fp
fm pp
(a) No Swelling Everywhere (b) Local Swelling (c) Uniform Swelling 0t t t
Fracture Compartment
Matrix
fp fp
fm pp 0
(d) No Swelling Everywhere (e) Local Swelling (f) Local Swelling 0t t t
fp
fm pp
Real Coal – Heterogeneous Real Coal – Heterogeneous Real Coal – Heterogeneous
Ideal Coal – Homogeneous Ideal Coal – Homogeneous Ideal Coal – Homogeneous
0ppm 0ppm
0ppm
0ppm
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-18
(1) Initial distributions of key parameters, including coal Young's modulus E ,
permeability 0k , porosity 0 , Langmuir strain constant L , and Langmuir
pressure constant Lp . If any or combination of them varies spatially, coal is
considered as initially heterogeneous.
(2) Swelling/shrinkage dependencies of these key parameters, including coal Young's
modulus E , Langmuir strain constant L , and Langmuir pressure constant Lp . If
any or combination of them varies spatially due to gas sorption, coal is considered as
swelling heterogeneous.
(3) Distribution of pore pressure p . The ultimate distribution of pore pressure is also
controlled by the boundary conditions. If this distribution is not uniform within coal
matrix, a uniform swelling within the matrix is also not achievable.
How to represent these heterogeneities more accurately is the key to reduce the difference
between the modelled coal permeability and the measured one. Based on the above review,
in the following section we represent the heterogeneity through spatial distributions of
Young's modulus and Langmuir strain constant.
6.3.3.1 Modelling scenarios for a heterogeneous matrix
We generalize changes in permeability that accompanies gas adsorption under conditions
of constant applied stress and for increments of applied gas pressure for fractures.
Specifically we explore the relations between coal transient (or dynamic) permeability and
equilibrium permeability, and how these relations are controlled both by the distributed
coal Young's modulus and Langmuir strain constant, and by the injection pressure. The
scenarios simulated in the model are listed in Table 6-3.
Table 6-3. Simulation scenarios for heterogeneous case
Simulation
scenario
Fracture pressure
fp (MPa)
Young's modulus
0E
Langmuir strain
constant L (%)
Modelling
result
Case 1 2.0, 4.0, 6.0, 8.0, 10
3.0 Figure 6-9(a)
Case 2 10.0
0.5, 0.75, 1.0, 1.25
3.0 Figure 6-9(b)
0E
0EMatrix
Fracture Wall
E
EMatrix
Fracture Wall
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-19
Case 3 10.0
1.0, 2.0, 3.0, 4.0 Figure 6-9(c)
Case 4 2.0, 4.0, 6.0, 8.0, 10 1.0
Figure 6-10(a)
Case 5 10.0 0.5, 0.75, 1.0, 1.25 Figure 6-10(b)
Case 6 10.0 1.0
1.0, 2.0, 3.0, 4.0
Figure 6-10(c)
Case 7 2.0, 4.0, 6.0, 8.0, 10
Figure 6-11(a)
Case 8 10.0
0.5, 0.75, 1.0, 1.25
Figure 6-11(b)
Case 9 10.0
1.0, 2.0, 3.0, 4.0
Figure 6-11(c)
6.3.3.2 Impacts of heterogeneous Young's modulus
In this section, the numerical results were summarized and divided into three different
groups. This division is based on the reasons for heterogeneity.
Cases 1 to 3 are for the heterogeneous coal represented by the spatial distribution of coal
Young's modulus. In all three cases, coal Young's modulus is considered to decrease
linearly from outer boundaries to the inner fracture walls, and its values at the outer
boundary and fracture wall are 0E and
0E respectively, where is defined as coal
Young's modulus softening coefficient. Langmuir strain constant is assumed as constant
throughout the whole simulated domain. As shown in Figure 6-9, the dynamic permeability
evolves from the initial rapid reduction, to recovery, and to a net reduction. The reason for
this net reduction is that as the localization extends to the outside boundary, as shown in
Figure 6-8(e), coal permeability recovers, but the even distribution of effective stress
induced strain is not achieved at the equilibrium state as the coal Young's modulus is
spatially related. More strain is expected near fracture wall as coal media is softer there than
outside shell, which is transferred to the reduction in fracture apertures, as illustrated in
Figure 6-8(f). This analysis explains why coal equilibrium permeability decreases with
increasing gas pore pressure and the Langmuir strain constants, as shown in Figures 6-9(a)
0E
0EMatrix
Fracture Wall
L
LMatrix
Fracture Wall
L
LMatrix
Fracture Wall
0E
0EMatrix
Fracture WallL
LMatrix
Fracture Wall
E
EMatrix
Fracture WallL
LMatrix
Fracture Wall
0E
0EMatrix
Fracture Wall
L
LMatrix
Fracture Wall
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-20
and 6-9(c), and increases with increasing coal Young's modulus changes, as shown in
Figure 6-9 (b).
6.3.3.3 Impacts of heterogeneous Langmuir strain constant
Cases 4 to 6 are for the heterogeneous coal represented by the spatial distribution of
Langmuir strain constant. This constant is considered to decrease linearly from the inner
fracture walls to outer boundaries, and its values at the outer boundary and fracture wall are
L and L respectively, where is swelling strain constant reduction factor. Coal
Young's modulus is assumed as constant throughout the simulations. Similar to the results
of cases 1 to 3, the dynamic permeability of these all cases also evolves from the initial
rapid reduction, to recovery, and to a net reduction, but a uniform swelling within coal
matrix is not achieved. When coal matrix swelling is localized near the fracture
compartment, this is to accommodate the swelling of a soft medium that is constrained
within a rigid outer shell. In this situation, coal sample can be considered to be constrained
from all directions, and coal matrix swelling is almost completely transferred to the
reduction in fracture apertures. Thus, most of the swelling strain contributes to coal
permeability change provided the fractures are much more compliant than coal matrix.
This analysis also explains why coal equilibrium permeability decreases at the equilibrium
state of gas transport. Modelling results are shown in Figures 6-10 (a) to (c).
6.3.3.4 Combined impacts of heterogeneous Langmuir strain and Young's modulus
Cases 7 to 9 are for the heterogeneous coal represented by the spatial distributions of both
coal Young's modulus and Langmuir strain constant. For these cases, coal Young's
modulus is considered to decrease linearly from outer boundaries to the inner fracture walls,
and its values at the outer boundary and fracture wall are 0E and 0E , respectively.
Langmuir strain constant is considered to decrease linearly from the inner fracture walls to
outer boundaries, and its values at the outer boundary and fracture wall are L and L ,
respectively. In all three cases, the dynamic permeability also evolves from the initial rapid
reduction, to recovery, and to a net reduction, but more permeability reduction was
observed, as shown in Figures 6-11 (a) to (c). For instance, the maximum equilibrium
permeability reduction is around 25%for cases 1 to 3, permeability decreases by 50% for
cases 4-6, and it further decrease by as much as 72% for cases 7 to 9, where both coal
Young's modulus and Langmuir strain constant are spatially dependent.
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-21
(a)
(b)
(c)
Figure 6-9. Evolutions of both dynamic permeability and equilibrium permeability for a
heterogeneous coal represented by the spatial distribution of coal Young's modulus under
the unconstrained swelling condition. (a) Influence of different pore pressures; (b)
influence of different modulus ratios; (c) influence different Langmuir strain constants.
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.0E+0 1.0E+1 1.0E+2 1.0E+3 1.0E+4 1.0E+5 1.0E+6
Dy
na
mic
pe
rme
ab
ilit
y r
ati
o (
k/k
0)
Time (s)
P=2.0MPa
P=4.0MPa
P=6.0MPa
P=8.0MPa
P=10.0MPa
Initial equilibriumFinal equilibrium
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
2 4 6 8 10
Fin
al e
qu
ilib
riu
m p
erm
ea
bilit
y r
ati
o
Pore pressure (MPa)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.0E+0 1.0E+1 1.0E+2 1.0E+3 1.0E+4 1.0E+5 1.0E+6
Dy
na
mic
pe
rme
ab
ilit
y r
ati
o (
k/k
0)
Time (s)
0.50 E0
0.75 E0
E0
1.25 E0
Initial equilibrium
Final equilibrium
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0.5 0.7 0.9 1.1 1.3
Fin
al e
qu
ilib
riu
m p
erm
ea
bilit
y r
ati
o
Young's modulus ratio (E/E0)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.0E+0 1.0E+1 1.0E+2 1.0E+3 1.0E+4 1.0E+5 1.0E+6
Dy
na
mic
pe
rme
ab
ilit
y r
ati
o (
k/k
0)
Time (s)
εs=1.0%
εs=2.0%
εs=3.0%
εs=4.0%
Initial equilibrium Final equilibrium
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1 1.5 2 2.5 3 3.5 4
Fin
al e
qu
ilib
riu
m p
erm
ea
bilit
y r
ati
o
Langmuir swelling strain constant (%)
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-22
(a)
(b)
(c)
Figure 6-10. Evolutions of both dynamic permeability and equilibrium permeability for a
heterogeneous coal represented by the spatial distribution of Langmuir strain constant
under the unconstrained swelling condition. (a) Influence of different pore pressures; (b)
influence of different modulus ratios; (c) influence of different initial values of Langmuir
strain constant.
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.0E+0 1.0E+1 1.0E+2 1.0E+3 1.0E+4 1.0E+5 1.0E+6
Dy
na
mic
pe
rme
ab
ilit
y r
ati
o (
k/k
0)
Time (s)
P=2.0MPa
P=4.0MPa
P=6.0MPa
P=8.0MPa
P=10.0MPa
Initial equilibrium
Final equilibrium
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
2 4 6 8 10
Fin
al e
qu
ilib
riu
m p
erm
ea
bilit
y r
ati
o
Pore pressure (MPa)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.0E+0 1.0E+1 1.0E+2 1.0E+3 1.0E+4 1.0E+5 1.0E+6
Dy
na
mic
pe
rme
ab
ilit
y r
ati
o (
k/k
0)
Time (s)
Initial equilibrium
Final equilibrium
0.50 E0
0.75 E0
E0
1.25 E0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0.5 0.7 0.9 1.1 1.3
Fin
al e
qu
ilib
riu
m p
erm
ea
bilit
y r
ati
o
Young's modulus ratio (E/E0)
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.0E+0 1.0E+1 1.0E+2 1.0E+3 1.0E+4 1.0E+5 1.0E+6
Dy
na
mic
pe
rme
ab
ilit
y r
ati
o (
k/k
0)
Time (s)
Initial equilibrium
Final equilibrium
εs=1.0%
εs=2.0%
εs=3.0%
εs=4.0%
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1 1.5 2 2.5 3 3.5 4
Fin
al e
qu
ilib
riu
m p
erm
ea
bilit
y r
ati
o
Langmuir swelling strain constant (%)
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-23
(a)
(b)
(c)
Figure 6-11. Evolutions of both dynamic permeability and equilibrium permeability for a
heterogeneous coal represented by the spatial distributions of both coal Young's modulus
and Langmuir strain constant under the unconstrained swelling condition. (a) Influence of
different pore pressures; (b) influence of different initial modulus; (c) influence of initial
Langmuir strain constants.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.0E+0 1.0E+1 1.0E+2 1.0E+3 1.0E+4 1.0E+5 1.0E+6
Dy
na
mic
pe
rme
ab
ilit
y r
ati
o (
k/k
0)
Time (s)
P=2.0MPa
P=4.0MPa
P=6.0MPa
P=8.0MPa
P=10.0MPa
Initial equilibrium
Final equilibrium
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
2 4 6 8 10
Fin
al e
qu
ilib
riu
m p
erm
ea
bilit
y r
ati
o
Pore pressure (MPa)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.0E+0 1.0E+1 1.0E+2 1.0E+3 1.0E+4 1.0E+5 1.0E+6
Dy
na
mic
pe
rme
ab
ilit
y r
ati
o (
k/k
0)
Time (s)
0.50 E0
0.75 E0
E0
1.25 E0
Initial equilibrium
Final equilibrium
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0.5 0.7 0.9 1.1 1.3
Fin
al e
qu
ilib
riu
m p
erm
ea
bilit
y r
ati
o
Young's modulus ratio (E/E0)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.0E+0 1.0E+1 1.0E+2 1.0E+3 1.0E+4 1.0E+5 1.0E+6
Dy
na
mic
pe
rme
ab
ilit
y r
ati
o (
k/k
0)
Time (s)
εs=1.0%
εs=2.0%
εs=3.0%
εs=4.0%
Initial equilibrium
Final equilibrium
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1 1.5 2 2.5 3 3.5 4
Fin
al e
qu
ilib
riu
m p
erm
ea
bilit
y r
ati
o
Langmuir swelling strain constant (%)
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-24
6.4 Verification with experimental data
In this section, two sets of experimental data monitored under stress controlled conditions
are used to verify our assumptions. The first experimental data comparison is based on the
work of Harpalani and Chen (1997), where the confining stress was kept constant
throughout the whole measurements. CH4 was used and permeability variation with
decreasing gas pressures (from 6.2MPa to 0.62MPa) was measured. The data for the second
match is from the work of Lin et al. (2008). Permeability of dry composite coal core to CH4
was measured with constant effective stress of 300psi (2.07MPa).
Both homogeneous and heterogeneous models are used to match two sets of data. For the
heterogeneous model, it is assumed that both coal Young's modulus and Langmuir strain
constant are distributed spatially. The parameter values used in data matching are listed in
Table 6-4 and the comparison results are shown in Figures 6-12 and 6-13, respectively.
Numerical results show that coal permeability value with homogeneous assumption
increases with increasing gas pressure, which is opposite to experimental data, so it is
incapable to match experimental data, but the heterogeneous model with the spatial
distributions of both coal Young's modulus and Langmuir strain is capable to replicate the
phenomena, and matches experimental data reasonably well. Therefore, coal heterogeneity
can be the answer to permeability change even under unconstrained conditions.
Table 6-4. Parameters used for data match
Parameter Harpalani and Chen (1997) Lin et al. (2008)
Porosity, % 5.0 5.0
Matrix Permeability, m2 5.0×10-22 2.0×10-21
Viscosity, Pa*s 1.228×10-5 1.228×10-5
Young's modulus, GPa 1.45 1.45
Poisson's ratio 0.4 0.4
Maximum volumetric swelling strain 0.005 0.0087
Langmuir sorption constant, m3/kg 0.0132 0.0132
Langmuir pressure constant, MPa 3.96 1.97
Coal density, kg/m3 1500 1500
Young's modulus softening coefficient, 0.79 0.555
Swelling strain reduction ratio, 0.21 0.445
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-25
Figure 6-12. Measured variation in permeability with decreasing gas pressure (Harpalani
and Chen, 1997).
Figure 6-13. Permeability of dry composite coal core to CH4. Effective stress equals to
300psi (Lin et al., 2008).
0
0.2
0.4
0.6
0.8
1
1.2
0 1 2 3 4 5 6 7
Pe
rme
ab
ilit
y R
ati
o (
k/k
0)
Gas pressure (MPa)
Experimental Data
Heterogeneous Case
Homogeneous Case
Fluid: CH4
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
0 1 2 3 4 5 6 7
Pe
rme
ab
ilit
y R
ati
o (
k/k
0)
Gas pressure (MPa)
Experimental Data
Heterogeneous Case
Homogeneous Case
Fluid: CH4
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-26
6.5 Conclusions
The performance of current coal permeability models is evaluated against analytical
solutions for the two extreme cases of either unconstrained swelling or constrained
swelling. Constrained model predictions are apparently consistent with both typical
laboratory measurements and in-situ observations. However, this apparent consistency is
due to the mismatch between model boundary condition assumptions (constrained) and
experiment boundary condition (unconstrained). This conclusion demonstrates that current
permeability models are incapable to explain net reductions in coal permeability where
swelling is unconstrained.
With the inclusion of the heterogeneous distributions of coal physical and swelling
properties, a full coupling approach was applied to investigate coal permeability response
under the unconstrained swelling conditions. Based on our model results, the following
major conclusions were drawn:
Both homogeneous and heterogeneous models experience the swelling transition
from local swelling to macro swelling. At the initial stage of gas injection, matrix
swelling is localized within the vicinity of the fracture compartment. As the
injection continues, the swelling zone is widening further into the matrix and
becomes macro-swelling.
Coal permeability experiences a rapid reduction at the early stage, a switch in
behavior from permeability reduction to recovery is observed, and coal
permeability finally recovers until it reaches the final equilibrium permeability. For
the homogeneous model, the final equilibrium permeability is always higher than
the original value, but the opposite was obtained for the heterogeneous model.
With the heterogeneous distributions of coal physical and swelling properties, this
numerical model matches with the experimental data reasonably well, which
demonstrates that heterogeneity of coal properties can be the answer to coal
permeability reduction under unconstrained conditions.
6.6 Acknowledgements
This work was supported by the Western Australia CSIRO-University Postgraduate
Research Scholarship, National Research Flagship Energy Transformed Top-up
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-27
Scholarship, ConocoPhillips, and NIOSH under contract 200-2008-25702. These sources
of support are gratefully acknowledged.
6.7 References
Ates, Y., Barron, K., 1988. The effect of gas sorption on the strength of coal. Mining Science and Technology
6, 291-300.
Chen, Z., Liu, J., Elsworth, D., Connell, L., Pan, Z., 2009. Investigation of CO2 Injection Induced Coal-Gas
Interactions, 43rd U.S. Rock Mechanics Symposium & 4th U.S. - Canada Rock Mechanics Symposium.
American Rock Mechanics Association, Asheville, North Carolina.
Chen, Z., Liu, J., Elsworth, D., Connell, L.D., Pan, Z., 2010. Impact of CO2 injection and differential
deformation on CO2 injectivity under in-situ stress conditions. International Journal of Coal Geology 81,
97-108.
Chen, Z., Pan, Z., Liu, J., Connell, L.D., Elsworth, D., 2011. Effect of the effective stress coefficient and
sorption-induced strain on the evolution of coal permeability: Experimental observations. International
Journal of Greenhouse Gas Control 5, 1284-1293.
Cherian, B.V., Claugus, A.O., Dilli, M., Mata, D., Sitchler, J., Alatrach, S., Neumiller, J.L., Panjaitan, M.L.,
2010. An Integrated Single-Well Approach to Evaluating Completion Effectiveness and Reservoir
Properties in the Wind Dancer Field, Tight Gas Completions Conference. Society of Petroleum
Engineers, San Antonio, Texas, USA.
Chilingar, G.V., 1964. Relationship Between Porosity, Permeability, and Grain-Size Distribution of Sands and
Sandstones, in: Straaten, L.M.J.U.v. (Ed.), Developments in Sedimentology. Elsevier, pp. 71-75.
Clarkson, C.R., 2008. Case Study: Production Data and Pressure Transient Analysis of Horseshoe Canyon
CBM Wells, CIPC/SPE Gas Technology Symposium 2008 Joint Conference. Society of Petroleum
Engineers, Calgary, Alberta, Canada.
Connell, L.D., 2009. Coupled flow and geomechanical processes during gas production from coal seams.
International Journal of Coal Geology 79, 18-28.
Connell, L.D., Detournay, C., 2009. Coupled flow and geomechanical processes during enhanced coal seam
methane recovery through CO2 sequestration. International Journal of Coal Geology 77, 222-233.
Connell, L.D., Lu, M., Pan, Z., 2010. An analytical coal permeability model for tri-axial strain and stress
conditions. International Journal of Coal Geology 84, 103-114.
Cui, X., Bustin, R.M., 2005. Volumetric strain associated with methane desorption and its impact on coalbed
gas production from deep coal seams. AAPG Bulletin 89, 1181-1202.
Cui, X., Bustin, R.M., Chikatamarla, L., 2007. Adsorption-induced coal swelling and stress: Implications for
methane production and acid gas sequestration into coal seams. J. Geophys. Res. 112, B10202.
Day, S., Fry, R., Sakurovs, R., 2008. Swelling of Australian coals in supercritical CO2. International Journal of
Coal Geology 74, 41-52.
Douglas, B., 1984. Microscopic in-situ studies of the solvent-induced swelling of thin sections of coal. Fuel 63,
1324-1328.
Durucan, S., Edwards, J.S., 1986. The effects of stress and fracturing on permeability of coal. Mining Science
and Technology 3, 205-216.
French, D.C., Dieckman, S.L., Botto, R.E., 1993. Three-dimensional NMR microscopic imaging of coal
swelling in pyridine. Energy & Fuels 7, 90-96.
Gathitu, B.B., Chen, W.-Y., McClure, M., 2009. Effects of Coal Interaction with Supercritical CO2: Physical
Structure. Industrial & Engineering Chemistry Research 48, 5024-5034.
Gibbins, J.R., Beeley, T.J., Crelling, J.C., Scott, A.C., Skorupska, N.M., Williamson, J., 1999. Observations of
Heterogeneity in Large Pulverized Coal Particles. Energy & Fuels 13, 592-601.
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-28
Goodman, A.L., Favors, R.N., Hill, M.M., Larsen, J.W., 2005. Structure Changes in Pittsburgh No. 8 Coal
Caused by Sorption of CO2 Gas. Energy & Fuels 19, 1759-1760.
Gu, F., Chalaturnyk, J.J., 2005. Analysis of Coalbed Methane Production by Reservoir and Geomechanical
Coupling Simulation.
Gu, F., Chalaturnyk, R.J., 2006. Numerical Simulation of Stress and Strain Due to Gas Sorption/Desorption
and Their Effects on In Situ Permeability of Coalbeds.
Han, F., Busch, A., van Wageningen, N., Yang, J., Liu, Z., Krooss, B.M., 2010. Experimental study of gas and
water transport processes in the inter-cleat (matrix) system of coal: Anthracite from Qinshui Basin, China.
International Journal of Coal Geology 81, 128-138.
Harpalani, S., Chen, G., 1997. Influence of gas production induced volumetric strain on permeability of coal.
Geotechnical and Geological Engineering 15, 303-325.
Harpalani, S., Schraufnagel, R.A., 1990. Shrinkage of coal matrix with release of gas and its impact on
permeability of coal. Fuel 69, 551-556.
Hsieh, S.T., Duda, J.L., 1987. Probing coal structure with organic vapour sorption. Fuel 66, 170-178.
Izadi, G., Wang, S., Elsworth, D., Liu, J., Wu, Y., Pone, D., 2011. Permeability evolution of fluid-infiltrated
coal containing discrete fractures. International Journal of Coal Geology 85, 202-211.
John W, L., 2004. The effects of dissolved CO2 on coal structure and properties. International Journal of
Coal Geology 57, 63-70.
Karacan, C.Ö., 2003. Heterogeneous Sorption and Swelling in a Confined and Stressed Coal during CO2
Injection. Energy & Fuels 17, 1595-1608.
Karacan, C.Ö., 2007. Swelling-induced volumetric strains internal to a stressed coal associated with CO2
sorption. International Journal of Coal Geology 72, 209-220.
Karacan, C.O., Okandan, E., 2001. Adsorption and gas transport in coal microstructure: investigation and
evaluation by quantitative X-ray CT imaging. Fuel 80, 509-520.
Kiyama, T., Nishimoto, S., Fujioka, M., Xue, Z., Ishijima, Y., Pan, Z., Connell, L.D., 2011. Coal swelling
strain and permeability change with injecting liquid/supercritical CO2 and N2 at stress-constrained
conditions. International Journal of Coal Geology 85, 56-64.
Larsen, J.W., 2004. The effects of dissolved CO2 on coal structure and properties. International Journal of
Coal Geology 57, 63-70.
Levine, J.R., 1996. Model study of the influence of matrix shrinkage on absolute permeability of coal bed
reservoirs, in: Gayer, R., Harris, I. (Eds.), Coalbed Methane and Coal Geology. Geological Society Special
Publication No 109. London, pp. 197- 212.
Lin, W., Tang, G.-Q., Kovscek, A.R., 2008. Sorption-Induced Permeability Change of Coal During Gas-
Injection Processes. SPE Reservoir Evaluation & Engineering.
Liu, C.J., Wang, G.X., Sang, S.X., Rudolph, V., 2010. Changes in pore structure of anthracite coal associated
with CO2 sequestration process. Fuel 89, 2665-2672.
Liu, H.-H., Rutqvist, J., 2010. A New Coal-Permeability Model: Internal Swelling Stress and Fracture–Matrix
Interaction. Transport in Porous Media 82, 157-171.
Liu, J., Chen, Z., Elsworth, D., Miao, X., Mao, X., 2010a. Evaluation of stress-controlled coal swelling
processes. International Journal of Coal Geology 83, 446-455.
Liu, J., Chen, Z., Elsworth, D., Miao, X., Mao, X., 2010b. Linking gas-sorption induced changes in coal
permeability to directional strains through a modulus reduction ratio. International Journal of Coal
Geology 83, 21-30.
Liu, J., Chen, Z., Elsworth, D., Qu, H., Chen, D., 2011a. Interactions of multiple processes during CBM
extraction: A critical review. International Journal of Coal Geology 87, 175-189.
Liu, J., Wang, J., Chen, Z., Wang, S., Elsworth, D., Jiang, Y., 2011b. Impact of transition from local swelling
to macro swelling on the evolution of coal permeability. International Journal of Coal Geology 88, 31-40.
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-29
Maggs, F.A.P., 1946. The adsorption-swelling of several carbonaceous solids. Transactions of the Faraday
Society 42, B284-B288.
Manovic, V., Loncarevic, D., Tokalic, R., 2009. Particle-to-Particle Heterogeneous Nature of Coals—A Case
of Large Coal Particles. Energy Sources, Part A: Recovery, Utilization, and Environmental Effects 31,
427-437.
Mazumder, S., Wolf, K.H., 2008. Differential swelling and permeability change of coal in response to CO2
injection for ECBM. International Journal of Coal Geology 74, 123-138.
Mirzaeian, M., Hall, P.J., 2006. The Interactions of Coal with CO2 and Its Effects on Coal Structure. Energy
& Fuels 20, 2022-2027.
Mirzaeian, M., Hall, P.J., 2008. Thermodynamical studies of irreversible sorption of CO2 by Wyodak coal.
Journal of Chemistry and Chemical Engineering 27, 59-68.
Palmer, I., Mansoori, J., 1996. How Permeability Depends on Stress and Pore Pressure in Coalbeds: A New
Model, SPE Annual Technical Conference and Exhibition. Society of Petroleum Engineers, Inc., Denver,
Colorado.
Pan, Z., Connell, L.D., Camilleri, M., 2010a. Laboratory characterisation of coal reservoir permeability for
primary and enhanced coalbed methane recovery. International Journal of Coal Geology 82, 252-261.
Pan, Z., Connell, L.D., Camilleri, M., Connelly, L., 2010b. Effects of matrix moisture on gas diffusion and
flow in coal. Fuel 89, 3207-3217.
Pekot, L.J., Reeves, S.R., 2002. Modeling the effects of matrix shrinkage and differential swelling on coalbed
methane recovery and carbon sequestration. U.S. Department of Energy DE-FC26-00NT40924.
Pini, R., Ottiger, S., Burlini, L., Storti, G., Mazzotti, M., 2009. Role of adsorption and swelling on the
dynamics of gas injection in coal. J. Geophys. Res. 114, B04203.
Pone, J.D.N., Halleck, P.M., Mathews, J.P., 2010. 3D characterization of coal strains induced by compression,
carbon dioxide sorption, and desorption at in-situ stress conditions. International Journal of Coal
Geology 82, 262-268.
Ranjith, P.G., Jasinge, D., Choi, S.K., Mehic, M., Shannon, B., 2010. The effect of CO2 saturation on
mechanical properties of Australian black coal using acoustic emission. Fuel 89, 2110-2117.
Robertson, E.P., 2005. Modeling Permeability in Coal Using Sorption-Induced Strain Data, SPE Annual
Technical Conference and Exhibition. Society of Petroleum Engineers, Dallas, Texas.
Seidle, J.R., Huitt, L.G., 1995. Experimental Measurement of Coal Matrix Shrinkage Due to Gas Desorption
and Implications for Cleat Permeability Increases, International Meeting on Petroleum Engineering.
Society of Petroleum Engineers, Inc., Beijing, China.
Shi, J.Q., Durucan, S., 2004. Drawdown Induced Changes in Permeability of Coalbeds: A New Interpretation
of the Reservoir Response to Primary Recovery. Transport in Porous Media 56, 1-16.
Siriwardane, H., Haljasmaa, I., McLendon, R., Irdi, G., Soong, Y., Bromhal, G., 2009. Influence of carbon
dioxide on coal permeability determined by pressure transient methods. International Journal of Coal
Geology 77, 109-118.
Somerton, W.H., Söylemezoglu, I.M., Dudley, R.C., 1975. Effect of stress on permeability of coal.
International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts 12, 129-145.
Sparks, D.P., McLendon, T.H., Saulsberry, J.L., Lambert, S.W., 1995. The Effects of Stress on Coalbed
Reservoir Performance, Black Warrior Basin, U.S.A, SPE Annual Technical Conference and Exhibition.
1995 Copyright 1995, Society of Petroleum Engineers, Inc., Dallas, Texas.
Viete, D.R., Ranjith, P.G., 2006. The effect of CO2 on the geomechanical and permeability behaviour of
brown coal: Implications for coal seam CO2 sequestration. International Journal of Coal Geology 66, 204-
216.
Viete, D.R., Ranjith, P.G., 2007. The mechanical behaviour of coal with respect to CO2 sequestration in deep
coal seams. Fuel 86, 2667-2671.
School of Mechanical and Chemical Engineering
Chapter 6 The University of Western Australia
6-30
Wang, G.X., Massarotto, P., Rudolph, V., 2009. An improved permeability model of coal for coalbed
methane recovery and CO2 geosequestration. International Journal of Coal Geology 77, 127-136.
Wang, G.X., Wei, X.R., Wang, K., Massarotto, P., Rudolph, V., 2010. Sorption-induced swelling/shrinkage
and permeability of coal under stressed adsorption/desorption conditions. International Journal of Coal
Geology 83, 46-54.
Wang, S., Elsworth, D., Liu, J., 2011. Permeability evolution in fractured coal: The roles of fracture geometry
and water-content. International Journal of Coal Geology 87, 13-25.
White, C.M., Smith, D.H., Jones, K.L., Goodman, A.L., Jikich, S.A., LaCount, R.B., DuBose, S.B., Ozdemir,
E., Morsi, B.I., Schroeder, K.T., 2005. Sequestration of Carbon Dioxide in Coal with Enhanced Coalbed
Methane RecoveryA Review. Energy & Fuels 19, 659-724.
Wu, Y., Liu, J., Elsworth, D., Chen, Z., Connell, L., Pan, Z., 2010a. Dual poroelastic response of a coal seam
to CO2 injection. International Journal of Greenhouse Gas Control 4, 668-678.
Wu, Y., Liu, J., Elsworth, D., Miao, X., Mao, X., 2010b. Development of anisotropic permeability during
coalbed methane production. Journal of Natural Gas Science and Engineering 2, 197-210.
Young, G.B.C., McElhiney, J.E., Paul, G.W., McBane, R.A., 1991. An Analysis of Fruitland Coalbed Methane
Production, Cedar Hill Field, Northern San Juan Basin, SPE Annual Technical Conference and
Exhibition. 1991 Copyright 1991, Society of Petroleum Engineers, Inc., Dallas, Texas.
Zhang, H., Liu, J., Elsworth, D., 2008. How sorption-induced matrix deformation affects gas flow in coal
seams: A new FE model. International Journal of Rock Mechanics and Mining Sciences 45, 1226-1236.
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-1
Chapter 7
Impact of Various Parameters on the Production of Coalbed Methane
Zhongwei Chena, Akim Kabirb, Jishan Liua, Jianguo Wanga, Zhejun Panc, Luke D. Connellc
a School of Mechanical and Chemical Engineering, The University of Western Australia, WA 6009, Australia
b QGC, A BG Business Group, 275 George Street, Brisbane, QLD 4000, Australia
c CSIRO Earth Science and Resource Engineering, Private Bag 10, Clayton South, Victoria 3169, Australia
Abstract: Coalbed methane (CBM) reservoirs are naturally fractured formations, consisting
of fractures (cleats) and coal matrixes. The interactions between fractures and matrixes
present a great challenge for the forecast of CBM productions. Although these interactions
were represented in our prior studies through dual poroelastic models, the impact of gas
mass transfer between fractures and coal matrixes on the CBM production have not been
well understood. In this study, the mass exchange between two systems is defined as a
function of gas sorption time constant at the standard condition, coal matrix porosity, and
the gas pressure difference between matrix and fracture systems. Correspondingly, the gas
diffusivity in the matrix is defined as a function of the shape factor, the gas sorption time,
and the gas pressure. These relations are integrated into a fully coupled finite element
model of coal deformation, gas flow and transport in the matrix system, and gas flow in the
fracture system. The FE model represents important non-linear impacts of the effective
stress on CBM productions that cannot be recovered where mechanical influences are not
rigorously coupled with the complex interactions between fractures and matrixes.
The FE model was applied to quantify the impact of various parameters on the CBM
production. These variables include (1) Gas Desorption Time Constant; (2) Initial Fracture
Permeability; (3) Fracture Spacing; (4) Swelling Capacity; and (5) Desorption Capacity. Our
numerical results show that the peak magnitudes of the gas production rate increase with
the initial fracture permeability, the swelling capacity and the sorption capacity, and
decrease with the gas desorption time constant and the production pressure. These results
also show dramatic increase in gas rate with decreasing magnitudes of the fracture spacing.
The comparison of the contributions of the desorbed gas, and the free phase gas from
fracture and matrix systems to the gas production rate shows that the free phase gas from
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-2
the matrix system plays the dominant role at the early stage, but diminishes when the
adsorption phase gas takes over the dominant role. The FE model is also applied to match
the production data of a single gas well production from the Fruitland Coal of San Juan
Basin. Model results match the actual production data well. This successful match
demonstrates the potential capability of the FE model for the forecast of CBM productions.
Keywords: coalbed methane; dual poroeasticity; gas production rate; numerical modeling
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-3
7.1 Introduction
Coal seams are naturally fractured reservoirs, consisting of fractures (cleats) and coal
matrixes. The micro-pores within the coal matrix are the main storage space for gas, while
the micro-fractures, fissures, fractures and faults comprise the main conduits for gas
seepage and migration. It has been widely recognised that the permeability of coalbed
methane reservoirs changes with pressure depletion during production, and the shrinkage
of coal matrix due to desorption of gas tends to open cleat fractures and increase coal
permeability, which are competitive and work in opposite directions (Harpalani and Chen,
1997; Gorucu, 2005).
Different to flow in low-permeability low-porosity formations, such as fractured crystalline
rocks, gas flow in unconsolidated materials such as coal is essentially interstitial where flow
routes may be rather tortuous. The fractured coal seam comprises both permeable fractures
and matrix blocks. Dual porosity representations (Barrenblatt et al., 1960; Warren and Root,
1963) include the response of these two principal components only – release from storage
in the porous matrix and transport in the fractured network. Conversely, dual permeability
or multiple permeability models represent the porosity and permeability of all constituent
components (Bai et al., 1993) including the role of linear sorption (Bai et al., 1997) and of
multiple fluids (Douglas et al., 1991). Traditional flow models accommodate the transport
response as overlapping continua but neglect geomechanical effects, which is an important
process that needs to be coupled in the numerical study (Connell, 2009).
Conceptualizations include analytical models for dual porosity media, their numerical
implementation and models including the component constitutive response for dual and
multi-porous (Aifantis, 1977; Bai et al., 1993; Bai and Elsworth, 2000; Elsworth and Bai,
1992) media. Such models have been applied to represent the response of permeability
evolution (Ouyang and Elsworth, 1993; Liu and Elsworth, 1997) in deforming aquifers (Bai
and Elsworth, 1994; Liu et al., 1999) and reservoirs (Bai et al., 1995). These mentioned
models were developed primarily for the flows of slightly compressible liquids without
nonlinear desorption, thus not applicable to the flow of compressible fluids such as
methane (CH4) where gasses desorption is the dominant mechanism.
Gas sorption and dissolution may cause the coal matrix to swell and/or shrink. This may
change the specific surface areas and total macro-pore volume of the coal matrix. When the
methane is produced, permeability was computed as a function of reservoir pressure and
coal matrix shrinkage. A number of theoretical and empirical permeability models have
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-4
been proposed (Chen et al., 2009; Connell, 2009; Connell and Detournay, 2009; Cui and
Bustin, 2005; Gu and Chalaturnyk, 2005; Izadi et al., 2011; Liu and Requivst, 2010; Liu et
al., 2010; Pekot and Reeves, 2002; Palmer and Mansoori, 1998; Seidle and Huitt, 1995; Shi
and Durucan, 2004; Zhang et al., 2008). These studies have contributed to our
understanding of the possible causes of fracture permeability changes, and the gas
transport in matrix system has been primarily assumed to be a diffusion process dominated
by gas desorption process. The desorption models in coal seam are classified into two
general groups: equilibrium (pressure dependent) models and non-equilibrium (pressure
and time dependent) models (King and Ertekin, 1995). In equilibrium models, it is assumed
that desorption/diffusion process occurs rapidly enough (sorption time is assumed to be
zero) that the kinetics of the process can be ignored in gas flow transport formulations;
whereas in non-equilibrium models, kinetics of desorption/diffusion is accommodated in
the transport equations using Fick′s diffusion laws (Roadifer and Moore, 2009; Ziarani et al.,
2011).
The sorption time is an important parameter to evaluate gas transport, and significant
findings have been drawn so far. Its influence on the gas production for a CBM reservoir
with a vertical well was first reported by Remener et al. (1986). Non-equilibrium sorption
phenomena for single-phase gas flow in coal seams were further investigated by Anbarci
and Ertekin (1990). Study by Ried et al. (1992) shows that sorption time is a less important
parameter for late-time production data history matching compared to other parameters
such as permeability, initial pressure and drainage area. Clarkson et al. (2006) showed that
early-time production data with non-equilibrium sorption (large sorption time, 110 days)
could be suppressed with respect to that of equilibrium sorption model. However, several
points can be improved for these studies. Firstly, the contribution of free gas from matrix
system is generally neglected, but it is unclear to what extent it affects reservoir prediction
results. Secondly, the diffusion coefficient was also assumed to be constant, which is not
consistent with other studies (Bird et al., 1960; Lama and Nguyen, 1987; Kranz et al., 1990).
Moreover, although the importance of the sorption time on coalbed methane extraction
has been widely investigated, many researchers have not taken the geomechanical process
into consideration (Palmer, 2009; Reeves and Pekot, 2001; Shi and Duracan, 2004), or
primarily assumed the absolute permeability to be a constant during both gas water
production and dry gas production (Clarkson et al., 2007; Ziarani et al., 2011), which
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-5
missed an important phenomenon of coalbed reservoir (Connell, 2009; Gu and
Chalaturnyk, 2005; Liu et al., 2011).
Another important parameter that could bring influence on gas production is the shape
factor. The shape factor concept, originally introduced by Barenblatt in 1960 and later by
Warren and Root (1963), provides an elegant and powerful upscaling method for fractured
reservoir simulation. The shape factor determines the fluid matrix blocks and the
surrounding fractures when there is a difference in pressure between two systems. An
appropriate specification of the shape factor is therefore critical for accurate modelling.
Studies have been conducted to determine the transfer shape factor for slightly
compressible fluids in the fractured reservoirs (Coats, 1989; Lim and Aziz, 1995; Kazemi,
1969; Thomas et al., 1983; Ueda et al., 1989). The functionality of the fracture pressure as a
boundary to the matrix blocks may also have a significant effect on the value of the shape
factor (Hassanzadeh and Pooladi-Darvish, 2006). Similar phenomenon has been obtained
by Ranjbar and Hassanzadeh (2011), where the effect of fracture pressure depletion regime
on the shape factor for single-phase flow of a compressible fluid was investigated and
obtained that shape factor is a function of the imposed boundary condition and its
variability with time. Recently, Heinemann and Mittermeir (2012) derived the Kazemi et al.
(1992) shape factor on the fracture-matrix dual continuum, and concluded that the Kazemi
et al. formula is exact under pseudo-steady-state conditions within the dual continuum
mathematical concept of natural fractured dual porosity systems. These expressions of the
shape factors have been used to investigate the behaviour of fracture-matrix primarily for
conventional gas reservoirs (Lim and Aziz, 1995; Mora and Wattattenbarger, 2009;
Zimmerman et al., 1993), but very little work can be found in the literature regarding its
impact on unconventional gas reservoirs (e.g. CBM) with coupled influence of
geomechanical deformation and gas desorption.
Although these interactions were represented in our prior studies through dual poroelastic
models, the impact of gas mass transfer between fractures and coal matrixes on the CBM
production have not been well understood. In this study, the mass exchange between two
systems is defined as a function of gas sorption time constant at the standard condition,
coal matrix porosity, and the gas pressure difference between matrix and fracture systems.
Correspondingly, the gas diffusivity in the matrix is defined as a function of the shape
factor, the gas sorption time, and the gas pressure. These relations are integrated into a
fully coupled finite element model of coal deformation, gas flow and transport in the
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-6
matrix system and gas flow in the fracture system. The FE model represents important
non-linear impacts of the effective stress on CBM productions that cannot be recovered
where mechanical influences are not rigorously coupled with the complex interactions
between fractures and matrixes. The FE model is then applied to quantify the impact of
various parameters on the CBM production. Afterwards model verification with field
observation is carried out followed by conclusions.
7.2 Methodology
CBM extraction induced complex interactions between stress and sorptive chemistry exert
strong influence on the transport and sorptive properties of the coal. These include
influences on gas sorption and flow, coal deformation, porosity change and permeability
modification. We label this chain of reactions as ―coupled processes‖ implying that one
physical process affects the initiation and progress of another. The individual processes, in
the absence of full consideration of cross couplings, form the basis of very well-known
disciplines such as elasticity, hydrology and heat transfer. Therefore, the inclusion of cross
couplings is the key to rigorously formulate the behavior for coupled processes of coal-gas
interactions. The complexity of these interactions is reflected in the extensive suite of
cross-coupling relations among coal deformation, gas transport in matrix, and gas flow and
transport in fracture as illustrated in Figure 7-1. These relations include:
(1) The interaction between coal deformation and gas flow in the fracture is defined by
the change of effective stress induced by pore pressure change in the fracture system,
and affects the volumetric strain of coal.
(2) The interaction between coal deformation and gas flow in the fracture is also defined
the porosity change and the permeability change induced by the coal deformation.
(3) The interaction between coal deformation and gas flow in the matrix is defined by the
change of variable diffusivity induced by pressure depletion and coal shrinkage in the
matrix, and affects the change in volumetric strain of coal.
(4) The interaction between gas flow in the fracture and gas flow in the matrix is defined
by the gas mass transfer between the two systems.
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-7
Figure 7-1. Conceptual representation of coal-gas interactions.
In order to recover important non-linear responses due to effective stress effects,
geomechanical influence has to be rigorously coupled with the gas transport system. This
can be achieved through a full coupling approach, which is often the preferred method for
simulating multiple types of physics simultaneously since it should theoretically produce the
most realistic results (Zhang et al., 2008; Liu et al. 2011; Chen et al., 2009 and 2010; Wu et
al., 2010 and 2011). For this approach, a single set of equations (generally a large system of
non-linear coupled partial differential equations) incorporating all of the relevant physics is
solved simultaneously. In this study, a mass exchange function for compressible fluid
between coal matrix and fracture systems is adapted and implemented into our simulation
framework. The derivation processes have been listed in detail in the following sections.
7.2.1 Coal deformation
For all equations, traditional conventions are used: a comma followed by subscripts
denotes differentiation with respect to spatial coordinates and repeated indices in the same
expression imply summation over the range of the indices. Based on the above
assumptions, the strain-displacement relation is expressed as
x
y
• Volumetric Change
• Porosity Change
• Permeability Change
• Volumetric Change
• Porosity Change
• Permeability Change
• Mass Exchange
• Deformation Transformation
• Porosity Change
• Permeability Change
COAL SEAM
COAL MATRIXSwelling Component
Non-swelling ComponentCOAL FRACTURE
Fig.1 Conceptual representation of coal-gas interactions
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-8
)(2
1,, ijjiij uu (7-1)
where ij is the component of the total strain tensor and iu is the component of the
displacement. The equilibrium equation is as
0, ijij f (7-2)
where ij denotes the component of the total stress tensor and if denotes the component
of the body force.
The gas sorption-induced strain s is supposed not to produce any shear strain. Its effects
on all three normal components of strain are the same. The constitutive relation (7-1) for
the deformed coal seam becomes
ijsijf
f
ijmm
ijkkijij pK
pKKGG
33
)9
1
6
1(
2
1 (7-3)
where )1(2
E
G , )21(3
E
K ,
K
Km 1 ,
s
fK
K
K
K1 . G is the shear
modulus of the fissured coal, E is Young's modulus of the fissured coal and is
Poisson's ratio of fissured coal. kk is the mean stress, mp and fp are the pore pressures in
matrix and fracture system, respectively. K and K represent the bulk modulus of coal
without fissures and with fissures, respectively. sK denotes the bulk modulus of coal grains.
m and f are defined as the fluid pressure ratio factors for matrix and fracture systems,
respectively, compatible with Biot's coefficient (Biot, 1941; Wilson and Aifantis, 1982). ij
is called Kronecker's delta, 1 for ji and 0 for ji . From Equation (7-3), one obtains
sffmmv ppK
)(1
(7-4)
where 332211 v is the volumetric strain of coal matrix and 3/kk is the
mean compressive stress. Combination of Equations (7-1), (7-3) and (7-4) yields the
following Navier-type equation:
0
21,,2,,,,
iim
Lm
LLiffimmkikkki fp
pp
pKppu
GGu
(7-5)
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-9
This equation is the governing equation for coal deformation. The two variables, mp and
fp are linked through gas flow equation.
7.2.2 Coal permeability model
In the analysis of coal permeability, the fractured coal mass is treated as a discontinuous
medium comprising both matrix and fractures (cleats). The individual matrix blocks are
represented by cubes and may behave isotropically with regard to swelling/shrinkage,
thermal expansion, and mechanical deformability (Liu et al., 1999). The cleats are the three
orthogonal fracture sets and may also have different apertures and mechanical properties
ascribed to the different directions.
In this study, the following coal permeability model is used, which includes the coupling
impact from geomechanical deformation, gas pressure change and coal shrinkage induced
strain. For the 3D case with three orthogonal sets of fractures, coal directional permeability,
xk , yk , and zk are defined as follows (see Liu et al. (2010) for detail)
3
00
)1(31
s
s
vt
f
m
K
pR
k
k
(7-6)
where 0f is the initial fracture porosity at reference conditions.
For the 2D case with two orthogonal sets of fractures, coal permeability is defined as
follows
3
00
)1(21
s
s
vt
f
m
K
pR
k
k
(7-7)
7.2.3 Mass exchange function
The concept of treating fractured reservoir as a dual-porosity medium was introduced by
Warren and Root (1963). Whenever a pressure difference exists between the matrix and
the fractures, a fluid flow between the media will occur. The volume transfer rate per unit
bulk volume between the matrix and the fracture has the form (Lim and Aziz, 1995; Heel
et al., 2008):
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-10
fmm ppV
kq
(7-8)
where was defined as the shape factor for experimental samples. V denotes the volume
of matrix blocks.
Gas transport in coal takes place under two simultaneous and parallel processes, i.e., gas
diffusion in the matrix (micro-pores) and gas flow in the fracture (macro-pores). It is
common to estimate the diffusion process using desorption time, which is the time
required to desorb 63.2% of the initial gas volume during a whole core desorption test.
The relation between desorption time, 0 , and the diffusivity,
0D , is defined as
00
0
1
D
(7-9)
The sorption time is a lumped parameter accounting both for diffusion and desorption
time. 0D is the gas diffusivity measured at 15.273aT K, and 325.101ap kPa.
The gas diffusivity under in situ condition can be related to fluid and coal properties (Bird
et al., 1960; Lama and Nguyen, 1987) by:
msrm
mkD
1
1 (7-10)
wherem is the effective matrix porosity, is gas (or fluid) compressibility (1/MPa), r is
the coal bulk compressibility (1/MPa), and s
is coal matrix compressibility (1/MPa).
For an isothermal case, the following relationship can be obtained (Kranz et al., 1990):
a
m
p
pDD 0 (7-11)
Gas compressibility is generally much higher than r , to a first approximation, it can be
simplified into:
m
mm
m
m pkkD (7-12)
Combining with Equations (7-9) and (7-11), the above equation can be expressed as:
00
0
1
a
m
a
mm
ppD
k (7-13)
Substituting Equation (7-13) into Equation (7-8) gives
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-11
0000
fm
a
mfm
a
mpp
p
Vpp
p
Vq
(7-14)
The above equation accounts for the mass exchange between matrix and fracture systems during
gas extraction.
7.2.4 Shape factor for dual-permeability model
In their 1960 landmark paper, Barenblatt et al. introduced the shape factor concept to
model the (single-phase) fluid transfer between matrix and fractures (1960). The central
idea of Barenblatt et al. was not to study the behaviour of individual matrix blocks and
their surrounding fractures, but instead to introduce two abstract interacting media: one
medium, the ―matrix‖, in which the physical matrix blocks are lumped, and one medium,
the ―fractures‖, in which the fractures are lumped.
The shape factor depends on the type of model selected for dual-porosity formulation. The
Warren and Root′s (1963), Kazemi′s (1969) and Lim and Aziz′s (1995) models are the most
widely used ones, which are formulated as follows.
Warren and Root′s model:
2
24
L
nn (7-15)
where 3 2, 1, n is sets of normal parallel fractures, and is associated with different flow
geometries (slabs, columns and cubes, respectively) and L is fracture spacing.
Kazemi′s model:
222
1114
zyx LLL (7-16)
Lim and Aziz (1995) presented analytical solutions of pressure diffusion for draining into a
constant fracture pressure (boundary condition) as
222
2 111
zyx LLL (7-17)
For equal fracture spacing, is equal to 223 L for three sets of fractures. For one and
two sets of fractures, the values for are 2 and
22 , respectively.
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-12
In this paper, comparison among three shape factors is conducted to study the difference
of the coalbed methane production behaviour to each factor value under dual-permeability
model.
7.2.5 Gas flow equations
In a dual-permeability model, there are matrix/fracture, matrix/matrix, and
fracture/fracture connections: each matrix has connections with the matrix of all its
neighbouring blocks, and each fracture node has connections with the fracture nodes of all
its neighbouring blocks (Bai et al., 1995). The dual-porosity model can be regarded as a
subset of the dual-permeability model: in a case where there are no contacts between
adjacent matrix blocks.
Based on the dual permeability system, the mass balance equation for each system is
expressed as
sgfgf
f
sgmgmm
Qqt
M
Qqt
M
(7-18)
where M is the gas content, g is the gas density, g
q is the Darcy's velocity vector, t is the
time. sQ is the gas mass exchange term that represents the net mass addition of fluid to the
fracture system from the matrix blocks, and defined as a function of Equation (7-14). mk is
the permeability for matrix system. It is assumed that the gas sorption take places in the
matrix system only. Therefore, the gas contents in the matrix and the fracture are written
out as:
Lm
mLcgamgmm
pp
pVM
(7-19)
fgffM (7-20)
where the subscript m represents for matrix, f for fracture, g for gas, and c for coal.
is porosity, ga is the gas density at standard conditions,
LV represents the Langmuir
volume constant, and Lp is the Langmuir pressure.
Based on the mass conservation law, substituting Equations (7-6), (7-14), (7-19) and (7-20)
into gas flow equations of both systems, and simplifying each one it gives
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-13
t
p
pp
p
t
pR
pp
p
pVpp
kp
tK
Rp
pp
p
pVpp
kp
tpp
pVp
m
Lm
LLv
f
fm
fm
a
mmff
f
f
sf
mff
fm
a
mmmm
mm
mL
LLcoalatmm
2
0
000
00
20
)1(3
1)1(3
(7-21)
where is the dynamic viscosity of gas.
Based on the diffusivity value at reference pressure (atmosphere pressure for this example),
substituting Equation (7-13) into Equation (7-21) yields
t
p
pp
p
t
pR
pp
p
pVpp
kp
tK
Rp
pp
p
pVpp
pp
tpp
pVp
m
Lm
LLv
f
fm
fm
a
mmff
f
f
sf
mff
fm
a
mmmm
a
mm
mL
LLcoalatmm
2
0
000
00
20
)1(3
1)1(3
1
(7-22)
As shown in Equation (7-22), gas flow in matrix is now a reverse function of desorption
time and shape factor. They affect gas transport in both matrix and fracture systems
through exchange term.
7.2.6 Gas Production Rate
For coal seam gas reservoirs, the original gas in place (OGIP) is defined as
dv
pp
pV
p
p
p
pOGIP
Li
iLcm
a
im
a
i (7-23)
where three terms on the right hand side represent free phase gas in the matrix, free phase
gas in the fracture, and adsorbed gas in the coal matrix, respectively. When the reservoir
reaches the abandoned pressure,abp , the residual gas in place (RGIP) can be defined as
dv
pp
pV
p
p
p
pRGIP
Lw
abLcf
a
abm
a
ab (7-24)
The difference between the original gas in place (OGIP) and the residual gas in place
(RGIP) gives the accumulative gas production, pG , as defined
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-14
dv
pp
pV
p
p
p
pdv
pp
pV
p
p
p
pG
Lw
abLcf
a
abm
a
ab
Li
iLcm
a
im
a
iT
(7-25)
For a closed coal seam gas reservoir, TG is constant. The gas production at time, t , is
defined as
dv
pp
pV
p
p
p
pdv
pp
pV
p
p
p
ptG
Lm
mLcf
a
f
m
a
m
Li
iLcm
a
im
a
ip
(7-26)
where mp and
fp are the pressures for matrix and fracture system at the time t, respectively.
The gas production rate is defined as
dv
dt
dp
pp
pV
dt
d
p
p
dt
dp
pdt
dp
pdt
tGdm
Lm
LLcf
a
ff
a
fm
a
mp
2
(7-27)
Therefore, the production rate is controlled primarily by the matrix pressure change rate,
dtdpm.
The matrix pressure change rate is dependent on the pressure transfer between coal matrix
and fracture induced by the gas production. Prior to the gas production, the gas pressure in
the fracture system is equal to that in the coal matrix. We define this state as the initial
equilibrium state, ifm ppp . Coal permeability at this state is defined as the initial
equilibrium permeability, 0fk . At this equilibrium state, no shrinking takes place anywhere.
When coal seam gas is extracted, the free phase gas flows out from the fracture system and
the gas pressure in the fracture reduces. At this stage, an imbalance between fracture
pressure and matrix pressure is achieved. This pressure imbalance changes as the gas in the
matrix transfers into the coal fracture. The pore pressure in the matrix decreases and as a
consequence coal matrix shrinks due to both the matrix pore pressure decrease and the gas
desorption. When the imbalance between fracture pressure and matrix pressure vanishes
completely, the final equilibrium state is achieved. At the final equilibrium state, the matrix
pressure is equal to the fracture pressure, i.e., fm pp . Coal permeability at this state is
defined as the final equilibrium permeability. When the final equilibrium state is reached,
0dtdpmindicates that no gas is produced from the reservoir.
Based on the analysis above, a number of factors could contribute to coal seam gas
production rate. These include (1) Desorption time, ; (2) Shape factor, ; (3) fk -
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-15
Fracture permeability; (4) Fracture spacing, L ; (5) Coal matrix swelling capacity, s ;(6) Gas
sorption capacity, LV ; (7) BHP– Bottom hole pressure; and (8) Matrix and fracture
porosities. The controlling effects of these factors are evaluated in the following section.
7.3 Evaluation of gas production
In this section, the sensitivity study of reservoir behaviour with respect to different
parameters is simulated, particularly the values for the shape factor and desorption time.
7.3.1 Model description
The simulation model geometry is 750m by 750m with a production well located at the
centre of the model. The wellbore diameter is 21.6 cm, and the computational model is
shown in Figure 7-2. For the coal deformation model, lateral directions were constrained.
For the gas transport model, coal sample was initially saturated with gas with certain
pressure, and a constant production pressure was applied on the wellbore boundary for all
cases. Input parameters for this simulation were listed in Table 7-2.
Figure 7-2. Vertical view of the geometry of numerical model
Φ=21.60mm
L=750.00 m
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-16
Table 7-1. Parameter values for simulation model
Parameter Value Parameter Value
Density of coal, ρc (kg/m3) 1500 Initial gas pressure, P0, (psi) 700
Matrix porosity (%) 5.0 Langmuir pressure constant ( MPa) 2.76
Fracture porosity (%) 0.5 Coal swelling capacity (%) 2.0
Fracture permeability (mD) 10 Coal matrix sorption capacity, m3/ton 6.0
Coal Young's modulus (psi) 500 Young's modulus reduction ratio (mR ) 0.1
Coal Poisson's ratio (-) 0.4 Desorption time (day) 17
Gas viscosity (Pa·s) 12.28×10-6 Fluid pressure ratio factor (m
) 0.2
Matrix size (yx LL ) (cm) 5.0 Fluid pressure ratio factor ( f ) 0.7
Shape factor ratio (0 ) 0.05 Coal seam thickness (m) 6.0
7.3.2 Modelling strategy
A series of gas extraction conditions as listed in Table 7-2 was simulated to investigate the
responses of gas production rate. Simulation results were presented in terms of (1) the
impacts of desorption time, (2) the impacts of initial fracture permeability value, (3) the
impacts of fracture spacing value (shape factor value), (4) the impacts of swelling strain
capacity, (5) the impacts of production pressure; (6) the impacts of sorption capacity; (7)
the impacts of fracture porosity; and (8) the impacts of matrix porosity. The modelling
results were shown in Figures 7-3 to 7-13.
Table 7-2. List of simulation scenarios
where * represents base case.
Case Parameter Value
1 Desorption time (day) 5; 17*; 95
2 Initial fracture permeability (mD) 1.0; 10*; 50
3 Fracture spacing (cm) 1.0; 5.0*; 10
4 Swelling strain capacity (%) 1.0; 2.0*; 3.0
5 Coal matrix sorption capacity (m3/ton) 3.0; 6.0*; 12
6 Production pressure (psi) 100; 200*; 300, 400
7 Fracture porosity (%) 0.1; 0.5*; 1.0
8 Matrix porosity (%) 1.0; 5.0*; 10
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-17
7.3.3 Reservoir behaviour under constant total productive gas volume
As shown in Equation (7-23), the total gas volume can be produced is determined by the
initial reservoir pressure, porosities, coal gas content, and the abandoned reservoir pressure
while the gas production processes are determined primarily by the matrix pressure change
rate and the fracture pressure change rate. In this section, we investigate the evolution of
gas production rate and cumulative gas production where the amount of total gas volume
can be produced is fixed. The investigated factors include cases 1 to 5, as listed in Table 7-2.
Results are shown in Figures 7-3 through 7-7.
Impacts of desorption time: For this case, three different desorption time values were
used, varying from 5 days to 50 days. A smaller desorption time, such as 5 , represents
a quicker desorption process in which the adsorbed gas on the surface of micropores
desorbs more quickly. On the other hand, a sorption time of 50 days represents a very slow
desorption/diffusion transport process of gas. As shown in Figure 7-3, the increase of
desorption time has a reverse effect on the peak of gas rate. The higher the desorption time
(slower desorption process), the smaller the peak rate, and the longer it takes to reach the
peak. For instance, it takes 134 days for 5 case to reach the peak value of 571
Mscf/day, but they require 254 days and 376 days to reach the maximum gas rate of 386
Mscf/day and 228 Mscf/day for 17 and 50 , respectively. A significant increase of
cumulative gas production with decreasing desorption time has been observed within this
simulated period, as shown in Figure 7-3(b). It is also worth to note that the difference of
production rate was observed from the beginning of the production.
(a) (b)
Figure 7-3. Evolution of gas production rate and cumulative gas production under different
magnitudes of gas desorption time
0
100
200
300
400
500
600
0 500 1000 1500 2000 2500 3000
Pro
du
cti
on
ra
te (
Ms
cf/
d)
Time (day)
τ=5 days
τ=17 days
τ=50 days
0
100
200
300
400
500
600
0 500 1000 1500 2000 2500 3000
Cu
mu
lati
ve
pro
du
cti
on
(M
Ms
cf)
Time (day)
τ=5 days
τ=17 days
τ=50 days
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-18
Impacts of initial fracture permeability: Figure 7-4 shows the evolution of gas
production rate and cumulative production under different magnitudes of initial fracture
permeability. An increase from 1.0 mD to 50 mD significantly reduces the time taken to
reach peak value and improves production performance. When the permeability is 50 mD,
gas production rate only needs 80 days to reach the peak of 817 Mscf/day, but it increases
to 245 days when the permeability is 10 mD and its peak value dwindles to 385 Mscf/day.
The peak value has yet to appear in the studying period for 1.0 mD case.
In this work, coal permeability was considered to be a variable and defined as a function of
coal deformation, pore pressure change and shrinkage strain, but it is still not very clear
how much this will affect reservoir behaviour comparing with conventional simulations
(non-coupling model or constant permeability). In the following, comparison of both gas
production rate and cumulative production between conventional simulations and the fully
coupling one is shown in Figure 7-5. It shows that the maximum gas rate and final gas
recovery from coupling simulations are higher than those from non-coupling approach for
the same initial fracture permeability. This demonstrates that coal permeability increases
with gas depletion as the shrinkage strain increases the fracture opening. The comparison
results also indicate the importance to consider the permeability as a variable and choose
more realistic permeability value.
(a) (b)
Figure 7-4. Evolution of gas production rate and cumulative gas production under different
magnitudes of initial fracture permeability
0
100
200
300
400
500
600
700
800
900
0 500 1000 1500 2000 2500 3000
Pro
du
cti
on
ra
te (
Ms
cf/
d)
Time (day)
k= 1.0 mD
k= 10 mD
k= 50 mD
0
100
200
300
400
500
600
0 500 1000 1500 2000 2500 3000
Cu
mu
lati
ve
pro
du
cti
on
(M
Ms
cf)
Time (day)
k= 1.0 mD
k= 10 mD
k= 50 mD
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-19
(a) (b)
Figure 7-5. Comparison of evolution of gas production rate and cumulative gas production
between coupling permeability model and non-coupling one
Impacts of fracture spacing: Figure 7-6 shows that with the increase of fracture spacing,
the maximum gas production rate and total gas recovery dramatically decrease, and the
same trend for the time to reach the peak gas rate. This finding has been consistent with
other researchers (Gu and Chalaturnyk, 2005; Sawyer et al., 1987). When the fracture
spacing value is 10 cm, the gas production rate takes 380 days to reach the peak value of
158 Mscf/day, but this value jumps to 760 Mscf/day at the day of 39 when fracture spacing
is 1.0cm. The primary reason for this effect is that as the fracture spacing becomes larger,
the mass exchange between two systems become slower, as shown in Equation (7-24), and
in turn postpones gas transport capacity in matrix system.
(a) (b)
Figure 7-6. Evolution of gas production rate and cumulative gas production under different
magnitudes of fracture spacing
0
50
100
150
200
250
300
350
400
450
0 500 1000 1500 2000 2500 3000
Pro
du
cti
on
ra
te (
Ms
cf/
d)
Time (day)
Non-coupling Model
Coupling Model
0
100
200
300
400
500
600
0 500 1000 1500 2000 2500 3000
Cu
mu
lati
ve
pro
du
cti
on
(M
Ms
cf)
Time (day)
Non-coupling Model
Coupling Model
0
100
200
300
400
500
600
700
800
0 500 1000 1500 2000 2500 3000
Pro
du
cti
on
ra
te (
Ms
cf/
d)
Time (day)
L=1.0 cm
L=5.0 cm
L=10 cm
0
100
200
300
400
500
600
0 500 1000 1500 2000 2500 3000
Cu
mu
lati
ve
pro
du
cti
on
(M
Ms
cf)
Time (day)
L=1.0 cm
L=5.0 cm
L=10 cm
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-20
Impacts of swelling capacity: The results from coupling simulations explore the
influence of the coal matrix swelling/shrinkage capacity (shrinkage for this case) due to
methane desorption are shown in Figure 7-7. Results suggest that for a coal seam with a
higher matrix shrinkage capacity, the maximum gas rate and final gas recovery are higher as
compared with a coal seam reservoir having a smaller value of matrix shrinkage capacity.
This is expected since a higher shrinkage capacity will cause a larger shrinkage and in turn
bring more fracture opening, thus a more significant increase in permeability as illustrated
in Equation (7-6). Different from the cases of desorption time and initial permeability value,
at the beginning (e.g. before 300 days), both gas production rate and total gas recovery for
different swelling capacities are very close. This demonstrates that the early stage
production is dominated by effective stress change due to pressure drawdown. The
increasing peak rate with increasing swelling capacity illustrates that the desorption-induced
permeability increase is playing increasingly important role on gas production.. Similar
phenomenon has been observed by other researchers even based on different permeability
models or different coal geometry models (Robertson, 2005; Shi and Durucan, 2004).
(a) (b)
Figure 7-7. Evolution of gas production rate and cumulative gas production under different
magnitudes of swelling strain capacity
In summary, for all cases, total gas volume can be produced in the reservoir is kept
constant through controlling parameters, as shown in Equation (7-25). The gas production
rate is determined by the matrix gas pressure change rate, dtdpm , and the fracture gas
pressure change rate,
dtdp f . They are dependent on each other. At the initial stage of gas
production, both pressures change with production, but the fracture pressure changes
much faster than the matrix gas pressure. This creates the pressure difference, so gas
0
50
100
150
200
250
300
350
400
450
0 500 1000 1500 2000 2500 3000
Pro
du
cti
on
ra
te (
Ms
cf/
d)
Time (day)
εL=1.0 %
εL=2.0 %
εL=3.0 %
0
100
200
300
400
500
0 500 1000 1500 2000 2500 3000
Cu
mu
lati
ve
pro
du
cti
on
(M
Ms
cf)
Time (day)
εL=1.0 %
εL=2.0 %
εL=3.0 %
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-21
production rate increases as production progresses. When the fracture gas pressure
stabilizes, 0dtdp f, the gas production rate is determined primarily by the matrix gas
pressure change rate. Gas production behaves different to different values of parameters,
which is a combining outcome of different mechanism, as shown in Equations (7-5) and
(7-24).
7.3.4 Reservoir behaviour under variable total productive gas volume
This part is to investigate the evolution of both gas production rate and cumulative gas
production under variable total gas volume in the reservoir, which relies on the production
condition such as production pressure, or gas storage condition such as coal sorption
capacity, and matrix and fracture porosities. The behaviours of coal seam gas reservoir to
different controlling factors are shown in Figures 7-8 through 7-12.
Impacts of sorption capacity: The effects of sorption capacity (or Langmuir volume) on
coalbed methane production are shown in Figure 7-8. Numerical results show that both
maximum gas production rate and total gas production rise with increasing gas sorption
capacity. For instance, the peak rate increases from 342 Mscf/day for 3.0 m3/ton case to
433 Mscf/day for 12m3/ton case, and the total gas production climbs from 361 MMscf for
3.0 m3/ton case during the simulation period to 652 MMscf for the later case. The time to
reach the maximum gas rate also slightly increases with increasing gas sorption capacity.
This is due to an increase of productive gas volume with the increase of gas sorption
capacity, while the reservoir pressure is fixed. This brings a larger gas supply which can
maintain higher gas production rate, and in turn postpones the decline of gas production
rate. It is worth noting that reservoir behaviour at the early stage both gas production rate
and total gas production follow the same trend, which is similar to the case for different
swelling capacities. This demonstrates that free gas phase plays more important than
sorption gas phase at this stage, as illustrated in Equation (7-24), but the desorption gas
phase takes over the dominant role of gas production after around 300 days, as shown in
Figure 7-8.
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-22
(a) (b)
Figure 7-8. Evolution of gas production rate and cumulative gas production under different
magnitudes of coal sorption capacity
Impacts of production pressure: It is a common practice to achieve the required gas
production rate through controlling bottomhole pressure (BHP). This sensitivity study is to
simulate how the reservoir behaves with respect to different BHP values. Results from
Figure 7-9 show that both gas production rate and cumulative production increase with
reducing bottomhole pressures. The maximum gas rates in the order of increasing BHP
values are 425 Mscf/day, 384 Mscf/day, 321 Mscf/day and 236 Mscf/day, respectively.
The cumulative production declines from 558 MMscf for 100p psi case to 263 MMscf
for 400p psi case. The dramatic difference in both gas rate and the cumulative gas
production against different BHP values are observed from the very early stage of
production, but only a little difference of the time takes to reach maximum value has been
observed.
(a) (b)
Figure 7-9. Evolution of gas production rate and cumulative gas production under different
magnitudes of gas production pressure
0
50
100
150
200
250
300
350
400
450
500
0 500 1000 1500 2000 2500 3000
Pro
du
cti
on
ra
te (
Ms
cf/
d)
Time (day)
VL=3 m3/ton
VL=6 m3/ton
VL=12 m3/ton
0
100
200
300
400
500
600
700
0 500 1000 1500 2000 2500 3000
Cu
mu
lati
ve
pro
du
cti
on
(M
Ms
cf)
Time (day)
VL=3 m3/ton
VL=6 m3/ton
VL=12 m3/ton
0
50
100
150
200
250
300
350
400
450
500
0 500 1000 1500 2000 2500 3000
Pro
du
cti
on
ra
te (
Ms
cf/
d)
Time (day)
P=100 psi
P=200 psi
P=300 psi
P=400 psi
0
100
200
300
400
500
600
0 500 1000 1500 2000 2500 3000
Cu
mu
lati
ve
pro
du
cti
on
(M
Ms
cf)
Time (day)
P=100 psi
P=200 psi
P=300 psi
P=400 psi
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-23
Impacts of fracture porosity: Figure 7-10 shows the evolution of gas production
regarding different fracture porosities. Model results show that fracture porosity has
significant influence on gas production rate profile. For instance, when %0.1f , gas rate
reaches its peak of 442 Mscf/day at 130 days, the maximum gas rate reduces to 385
Mscf/day at 305 days when f decreases to 0.5%, while it rebounds to 412 Mscf/day at
508 days when fracture porosity further decreases to 0.1%, as illustrated in Figure 7-10(a).
This rebound can be explained from two competing mechanisms that fracture porosity
contributes in the flow equation: on the one hand, decreasing fracture porosity reduces free
gas phase volume, which reduces gas production rate, but on the other hand reducing
fracture porosity increases the sensitivity of coal fracture permeability, which means more
permeability increase is expected during gas depletion. This enhancement of fracture
permeability can bring increase in gas production rate, as explained for the impacts of
fracture permeability on reservoir performance in Figure 7-5. This explanation can also be
seen from Equation (7-6), and has been confirmed by other researchers as well (Harpalani
and Chen, 1995, Gu and Chalaturnyk, 2005; Palmer, 2010). The gain or lost of gas
production rate depends on the competing processes of both mechanisms. Our numerical
results show that the influence of free gas volume change plays more significant role than
fracture permeability does for both %0.1f and %5.0f case, but it has been the
opposite for %1.0f case. Meanwhile, the comparison of cumulative production for
three cases indicates that the influence of the fracture porosity is only significant for the
early stage of gas production, as plotted in Figure 7-10(b).
(a) (b)
Figure 7-10. Evolution of gas production rate and cumulative gas production under
different magnitudes of fracture porosity
0
50
100
150
200
250
300
350
400
450
500
0 500 1000 1500 2000 2500 3000
Pro
du
cti
on
ra
te (
Ms
cf/
d)
Time (day)
Φf= 0.1%
Φf= 0.5%
Φf= 1.0%
0
100
200
300
400
500
600
0 500 1000 1500 2000 2500 3000
Cu
mu
lati
ve
pro
du
cti
on
(M
Ms
cf)
Time (day)
Φf= 0.1%
Φf= 0.5%
Φf= 1.0%
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-24
Impacts of matrix porosity: The performance of coal gas reservoir to different
magnitudes of matrix porosity is illustrated in Figure 7-11. An increase in both gas rate and
cumulative production with increasing matrix porosity is obtained, but a slight decrease for
the time to reach peak value is also observed. When %0.1m , the maximum gas rate is
around 150 Mscf/day, and it quickly jumps to 385 Mscf/day and 565 Mscf/day for
%0.5m and %10m cases, respectively.
However, this finding is contrary to the general belief that gas is stored primarily by
sorption into the coal and accounts for 98% of the gas within a coal seam (Gray, 1987),
and as a result the influence of free gas term in matrix porosity on gas produced has been
simply dismissed in many studies, particularly for the studies using Fick‘s law (Busch et al.,
2004; Clarkson and Bustin, 1999; King et al., 1986; Mora and Wattattenbarger, 2009; Shi
and Durucan, 2008). Equation (7-24) illustrates that two ways the matrix porosity can
contribute for gas transport in matrix system: increases free gas volume and enhances mass
exchange rate between two systems. Figure 7-12 illustrates the comparison of two different
numerical models: one couples the free gas influence of matrix system, and the other does
not consider the matrix free gas influence. Our numerical results show that model without
free gas influence could significantly overestimate gas production behaviour.
Comparison of the transient contribution of the desorbed gas, and the free phase gas in
both matrix and fracture systems to gas production rate at each moment for %0.5m
and %10m cases are shown in Figure 7-13 (a, c). It needs to be pointed out that the
ratio of the contribution for each system is calculated based on the change rate of each
system at each transient moment over the gas rate for the same moment. The detailed
explanation of the transient contribution calculation is listed in Table 7-3. Figure 7-13 (a, c)
shows that the free gas in matrix porosity plays the dominant role for gas transport at the
early stage, with maximum portion of around 48% for %0.5m case, which means 48%
of gas volume production at that moment is from matrix system (the source of this gas may
from desorption, but it is stored in matrix system at that moment), and increases to 66%
for %10m case. But the transient contribution of free gas in matrix porosity declines
with gas extraction and eventually adsorption phase gas takes over the dominant role. The
desorption gas and free phase gas in matrix system contribute to 71% and 28% of the gas
production rate for %0.5m case at the final investigation period, and their values
change into 55.8% and 43.5% for %10m case.
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-25
Comparison of the transient contributions of each part on total gas accumulation for two
different matrix porosities is plotted in Figure 7-13 (b, d). It can be seen that desorption
gas plays the increasingly dominant role on total gas production. For instance, for
%0.5m case, at the beginning of gas production, adsorption gas phase and free gas of
matrix system contribute 68.3% and 28.8% of total gas production, respectively, but these
values change into 78.2% and 19.4% at the end of simulation period. Similar trend is also
observed for %10m case with 53.1% and 44.7% at the beginning, and become 65.8%
and 33.2% at the end of simulation period.
From the above analysis we can see that although a great percentage of gas is stored in
adsorbed phase in coal seam gas reservoir, the real contribution from adsorbed phase at gas
transient flow process may be much lower than expected depending on reservoirs, and it
could overestimate reservoir behaviour is free gas from matrix system is dismissed in the
numerical model. Therefore, it could be more reasonable to consider the transient
contribution from matrix porosity when gas transport process is simulated to achieve more
realistic results.
(a) (b)
Figure 7-11. Evolution of gas production rate and cumulative gas production under
different magnitudes of matrix porosity.
0
100
200
300
400
500
600
700
0 500 1000 1500 2000 2500 3000
Pro
du
cti
on
ra
te (
Ms
cf/
d)
Time (day)
Φm= 1.0%
Φm= 5.0%
Φm= 10%
0
100
200
300
400
500
600
700
800
0 500 1000 1500 2000 2500 3000
Cu
mu
lati
ve
pro
du
cti
on
(M
Ms
cf)
Time (day)
Φm= 1.0%
Φm= 5.0%
Φm= 10%
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-26
Figure 7-12. Comparison of gas produciton between numerical models with fress gas
influence and without
Table 7-3. Explanation of the transient contribution calculation
Time Free Gas in
Matrix Free Gas in Fracture Adsorbed Gas Total
0T 0,mV 0,fV 0,adV 0V
ttT ttmV , ttfV , ttadV , ttV
tT tmV , tfV , tadV , tV
Transient
contribution
on gas rate
ttt
tmttm
VV
VV
,,
ttt
tfttf
VV
VV
,,
ttt
tadttad
VV
VV
,, -
Transient
contribution on gas
production t
tmm
VV
VV
0
,0,
t
tff
VV
VV
0
,0, t
tadad
VV
VV
0
,0, -
0
100
200
300
400
500
600
0
100
200
300
400
500
600
0 500 1000 1500 2000 2500 3000
Pro
du
cti
on
ra
te (
Ms
cf/
d)
Time (day)
Φm= 5.0%
Gas rate (with free gas)
Cu
mu
lati
ve
pro
du
cti
on
(MM
sc
f)
Gas rate (without free gas)
Production (with free gas)
Production(without free gas)
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-27
(a) (b)
(c) (d)
Figure 7-13. Transient contributions of coal matrix system, fracture system, and desorption
process to gas production rate and cumulative gas production for two difference matrix
porosity cases
7.4 Field application of dual-permeability model
The comparison between the numerical results from dual-permeability model and the field
observations is qualitatively made to check the feasibility of the developed numerical model.
This example is from a single gas phase well production from the Fruitland Coal of San
Juan Basin, which has been presented in other studies (Clarkson and McGovern, 2005;
Clarkson et al., 2006; Clarkson et al., 2007).
Critical data collected from the pilot well included reservoir pressure (500 psia), original gas
in place (OGIP) (≈3.0 Bscf), wellbore diameter (20 cm), and production data. The drainage
area for the well is 320 acres. For the history-matching process, as a lot of parameters are
not available from literature, the OGIP value was used to calculate gas contents and
adsorption isotherms. Coal physical properties, such as Young‘s modulus and Poisson‘s
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 500 1000 1500 2000 2500 3000
Tra
ns
ien
t c
on
trib
utio
n o
n g
as
p
rod
uc
tio
n ra
te (
%)
Time (day)
From matrix system
From fracture system
From desorption
Φm= 5.0%
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 500 1000 1500 2000 2500 3000
Tra
ns
ien
t c
on
trib
utio
n o
n g
as
p
rod
uc
tio
n (%
)
Time (day)
From matrix system
From fracture system
From desorption
Φm= 5.0%
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 500 1000 1500 2000 2500 3000
Tra
ns
ien
t c
on
trib
utio
n o
n g
as
p
rod
uc
tio
n ra
te (
%)
Time (day)
From matrix system
From fracture system
From desorption
Φm= 10%
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 500 1000 1500 2000 2500 3000
Tra
ns
ien
t c
on
trib
utio
n o
n g
as
p
rod
uc
tio
n (%
)
Time (day)
From matrix system
From fracture system
From desorption
Φm= 10%
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-28
ratio, are cited from experimental measurements from other publications (Hager and Jones,
2001; Mavor and Vaughn, 1998; Siriwardane et al., 2009). Key data associated with geologic
and reservoir simulation models are listed in Table 7-3. In this matching model, the
constrained boundary conditions are applied and the monitored flowing bottomhole
pressure is used to simulate gas production rate and to compare with field data.
Comparison between numerical results of both dual-permeability model and single
permeability mode and field observations is presented in Figures 7-14. The single
permeability model refers to the model where the free gas in the matrix system is not
accounted, and the same parameters are used for both models.
Table 7-4. Parameter values for field data match
Parameter Value Parameter Value
Density of coal, ρc (kg/m3) 1300 Initial gas pressure, P0, (psi) 500
Matrix porosity (%) 4.4 Langmuir pressure constant ( MPa) 5.67
Fracture porosity (%) 0.5 Coal swelling capacity (%) 2.0
Fracture permeability (md) 30 Coal matrix sorption capacity, m3/ton 4.2
Coal Young's modulus (psi) 521 Young's modulus reduction ratio (mR ) 0.1
Coal Poisson's ratio (-) 0.32 Desorption time (day) 8
Gas viscosity 12.28×10-6 Fluid pressure ratio factor ( m ) 0.2
Matrix size (yx LL ) (cm) 5.0 Fluid pressure ratio factor ( f ) 0.7
Shape factor (m-2) Kazemi′s model Coal seam thickness (m) 14
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-29
Figure 7-14. Comparison between numerical results and field observations.
Comparison results from Figure 7-14 show that the dual-permeability model matches the
field data reasonably well, indicating that dual-permeability is capable to replicate the real
reservoir behavior. It is worth noting that the increase in gas rate at the early stage until
reaching the peak flow rate demonstrates the enhancing contribution from the mass
exchange between matrix and fracture systems, which is dominated by gas desorption from
coal matrix. The decline of gas rate in the late stage illustrates the gas desorption rate
decreases with pressure drawdown.
7.5 Conclusion
In this study, the dual-porosity/dual-permeability geometry was incorporated to conduct a
finite element numerical model, which combines with the complex geomechanical
deformation, gas flow and transport in the matrix system, and gas flow and transport in the
fracture system together with variable permeability. In this model, the mass exchange
between matrix and fracture systems was given as a function of sorption time, matrix
porosity, matrix particle volume, and pressure difference between matrix and fracture
systems.
Based on the modelling results and data comparison with field observations, the following
conclusions can be drawn:
0
100
200
300
400
500
600
700
10
100
1000
0 1000 2000 3000 4000
Ga
s p
rod
uc
tio
n r
ate
(M
sc
f/d
)
Time (day)
Gas rate
Dual-permeability model
Single permeability model
Flowing BHP
Input BHP data
Flo
win
gb
ott
om
ho
le p
res
su
re (
Ps
ia)
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-30
(1) Sorption time has a reverse effect on the peak of gas rate. The higher the sorption
time, the lower the peak of production rate, and the longer it takes to reach the
peak. The cumulative gas production increases with decreasing sorption time
values.
(2) Results from dual-permeability model show that fracture spacing has significant
influence on gas production. Higher gas rate is observed for smaller fracture
spacing cases, while shape factor model has little influence on compressible gas
production, which has been significantly different from the behavior of
incompressible or slightly compressible reservoirs.
(3) The initial fracture permeability, gas sorption capacity (or Langmuir volume) and
swelling capacity have similar impacts on gas production performance. The gas
production enhances with increasing values of these parameters as well as the time
to reach the maximum gas rate.
(4) Comparison of the transient contributions of the desorbed gas, and the free phase
gas from both two systems to gas production rate at each moment show that the
free gas from matrix plays the dominant role for gas transport at the early stage,
but declines with gas extraction and eventually desorbed phase gas takes over the
dominant role. For instance, the desorbed gas and free phase gas in matrix system
contribute to 71% and 28% of the gas production rate for %0.5m case at the
final investigation period, and their values change into 55.8% and 43.5% for
%10m case. However, the desorbed gas phase plays the increasingly dominant
role on total gas production, but the weight decreases with increasing matrix
porosities. For example, the desorbed gas contributes 78.2% of total gas
production for %5m case at the end of simulation period, but this weight
reduces to 65.8% for %10m case.
7.6 Acknowledgements
This work was supported by the Western Australia CSIRO-University Postgraduate
Research Scholarship, National Research Flagship Energy Transformed Top-up
Scholarship. These sources of support are gratefully acknowledged.
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-31
7.7 References
Aifantis, E.C., 1977. Introducing a multi-porous medium. Developments in Mechanica 37, 265–296.
Anbarci, K., Ertekin, T., 1990. A Comprehensive Study of Pressure Transient Analysis With Sorption
Phenomena for Single-Phase Gas Flow in Coal Seams, SPE Annual Technical Conference and
Exhibition. 1990 Copyright 1990, Society of Petroleum Engineers Inc., New Orleans, Louisiana.
Bai, M., Elsworth, D., 1994. Modeling of subsidence and stress-dependent hydraulic conductivity for intact
and fractured porous media. Rock Mechanics and Rock Engineering 27, 209-234.
Bai, M., Elsworth, D., 2000. Coupled Processes in Subsurface Deformation, Flow and Transport. ASCE
Press.
Bai, M., Elsworth, D., Inyang, H.I., Roegiers, J.C., 1997. Modeling Contaminant Migration with Linear
Sorption in Strongly Heterogeneous Media. J Environ Eng (New York) 123, 1116-1125.
Bai, M., Elsworth, D., Roegiers, J.-C., 1993. Multiporosity/multipermeability approach to the simulation of
naturally fractured reservoirs. Water Resour. Res. 29, 1621-1633.
Bai, M., Roegiers, J.-C., Elsworth, D., 1995. Poromechanical response of fractured-porous rock masses.
Journal of Petroleum Science and Engineering 13, 155-168.
Barenblatt, G.I., 1960. Basic concepts in the theory of seepage of homogeneous liquids in fissured rocks.
Journal of applied mathematics and mechanics 24, 1286. Biot, M., 1941. General Theory of Three-
Dimensional Consolidation. Journal of Applied Physics 12, 155-164.
Biot, M., 1941. General Theory of Three-Dimensional Consolidation. Journal of Applied Physics 12, 155-164.
Bird, R.B., Stewart, W.E., Lightfoot, E.W., 1960. Transport Phenomena. Wiley, New York.
Busch, A., Gensterblum, Y., Krooss, B.M., Littke, R., 2004. Methane and carbon dioxide adsorption–
diffusion experiments on coal: upscaling and modeling. International Journal of Coal Geology 60,
151-168.
Chen, Z., Liu, J., Elsworth, D., Connell, L., Pan, Z., 2009. Investigation of CO2 Injection Induced Coal-Gas
Interactions, 43rd U.S. Rock Mechanics Symposium & 4th U.S. - Canada Rock Mechanics Symposium.
American Rock Mechanics Association, Asheville, North Carolina.
Chen, Z., Liu, J., Elsworth, D., Connell, L.D., Pan, Z., 2010. Impact of CO2 injection and differential
deformation on CO2 injectivity under in-situ stress conditions. International Journal of Coal Geology
81, 97-108.
Clarkson, C.R., Bustin, R.M., 1999. The effect of pore structure and gas pressure upon the transport
properties of coal: a laboratory and modeling study. 2. Adsorption rate modeling. Fuel 78, 1345-1362.
Clarkson, C.R., Bustin, R.M., Seidle, J.P., 2006. Production Data Analysis of Single-Phase (Gas) CBM Wells,
SPE Gas Technology Symposium. Society of Petroleum Engineers, Calgary, Alberta, Canada.
Clarkson, C.R., Bustin, R.M., Seidle, J.P., 2007. Production-Data Analysis of Single-Phase (Gas) Coalbed-
Methane Wells. SPE Reservoir Evaluation & Engineering 10.
Clarkson, C.R., McGovern, J.M., 2005. Optimization of Coalbed-Methane-Reservoir Exploration and
Development Strategies Through Integration of Simulation and Economics. SPE Reservoir
Evaluation & Engineering.
Coats, K.H., 1989. Implicit Compositional Simulation of Single-Porosity and Dual-Porosity Reservoirs, SPE
Symposium on Reservoir Simulation. 1989 Copyright 1989, Society of Petroleum Engineers, Houston,
Texas.
Connell, L.D., 2009. Coupled flow and geomechanical processes during gas production from coal seams.
International Journal of Coal Geology 79, 18-28.
Connell, L.D., Detournay, C., 2009. Coupled flow and geomechanical processes during enhanced coal seam
methane recovery through CO2 sequestration. International Journal of Coal Geology 77, 222-233.
Cui, X., Bustin, R.M., 2005. Volumetric strain associated with methane desorption and its impact on coalbed
gas production from deep coal seams. AAPG Bulletin 89, 1181-1202.
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-32
Douglas Jr, J., Hensley, J.L., Arbogast, T., 1991. A dual-porosity model for waterflooding in naturally
fractured reservoirs. Computer Methods in Applied Mechanics and Engineering 87, 157-174.
Elsworth, D., Bai, M., 1992. Flow-Deformation Response of Dual-Porosity Media. Journal of Geotechnical
Engineering 118, 107-124.
Gorucu, F.B., Jikich, S.A., Bromhal, G.S., Sams, W.N., Ertekin, T., Smith, D.H., 2005. Matrix Shrinkage and
Swelling Effects on Economics of Enhanced Coalbed Methane Production and CO2 Sequestration in
Coal, SPE Eastern Regional Meeting. Society of Petroleum Engineers, Morgantown, West Virginia.
Gray, I., 1987. Reservoir Engineering in Coal Seams: Part 1-The Physical Process of Gas Storage and
Movement in Coal Seams. SPE Reservoir Engineering.
Gu, F., Chalaturnyk, R.J., 2005. Sensitivity Study of Coalbed Methane Production With Reservoir and
Geomechanic Coupling Simulation.
Hager, C.J., Jones, J.R., 2001. Analyzing Flowing Production Data with Standard Pressure Transient Methods,
SPE Rocky Mountain Petroleum Technology Conference. Copyright 2001, Society of Petroleum
Engineers Inc., Keystone, Colorado.
Harpalani, S., Chen, G., 1995. Estimation of changes in fracture porosity of coal with gas emission. Fuel 74,
1491-1498.
Harpalani, S., Chen, G., 1997. Influence of gas production induced volumetric strain on permeability of coal.
Geotechnical and Geological Engineering 15, 303-325.
Hassanzadeh, H., Pooladi-Darvish, M., 2006. Effects of Fracture Boundary Conditions on Matrix-fracture
Transfer Shape Factor. Transport in Porous Media 64, 51-71.
Heel, A.P.v., Boerrigter, P.M., Dorp, J.J.v., 2008. Thermal and Hydraulic Matrix-Fracture Interaction in Dual-
Permeability Simulation. SPE Reservoir Evaluation & Engineering.
Heinemann, Z., Mittermeir, G., 2012. Derivation of the Kazemi–Gilman–Elsharkawy Generalized Dual
Porosity Shape Factor. Transport in Porous Media 91, 123-132.
Izadi, G., Wang, S., Elsworth, D., Liu, J., Wu, Y., Pone, D., 2011. Permeability evolution of fluid-infiltrated
coal containing discrete fractures. International Journal of Coal Geology 85, 202-211.
Kazemi, H., 1969. Pressure Transient Analysis of Naturally Fractured Reservoirs with Uniform Fracture
Distribution.
Kazemi, H., Gilman, J.R., Elsharkawy, A.M., 1992. Analytical and Numerical Solution of Oil Recovery From
Fractured Reservoirs With Empirical Transfer Functions (includes associated papers 25528 and
25818). SPE Reservoir Engineering.
King, G.R., Ertekin, T., 1995. State-of-the-Art Modeling for Unconventional Gas Recovery, Part II:Recent
Developments, Low Permeability Reservoirs Symposium. Copyright 1995, Society of Petroleum
Engineers Inc., Denver, Colorado.
King, G.R., Ertekin, T., Schwerer, F.C., 1986. Numerical Simulation of the Transient Behavior of Coal-Seam
Degasification Wells. SPE Formation Evaluation.
Kranz, R.L., Saltzman, J.S., Blacic, J.D., 1990. Hydraulic diffusivity measurements on laboratory rock samples
using an oscillating pore pressure method. International Journal of Rock Mechanics and Mining
Sciences & Geomechanics Abstracts 27, 345-352.
Lama, R.D., Nguyen, V.U., 1987. A model for determination of methane flow parameters in coal from
desorp- tion tests, the Twentieth International Symposium on the Application of Computers and
Mathematics in the Mineral Industries. SAIMM, Johannesburg, pp. 275-282.
Lim, K.T., Aziz, K., 1995. Matrix-fracture transfer shape factors for dual-porosity simulators. Journal of
Petroleum Science and Engineering 13, 169-178.
Liu, H.-H., Rutqvist, J., 2010. A New Coal-Permeability Model: Internal Swelling Stress and Fracture–Matrix
Interaction. Transport in Porous Media 82, 157-171.
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-33
Liu, J., Chen, Z., Elsworth, D., Miao, X., Mao, X., 2010. Linking gas-sorption induced changes in coal
permeability to directional strains through a modulus reduction ratio. International Journal of Coal
Geology 83, 21-30.
Liu, J., Chen, Z., Elsworth, D., Qu, H., Chen, D., 2011. Interactions of multiple processes during CBM
extraction: A critical review. International Journal of Coal Geology 87, 175-189.
Liu, J., Elsworth, D., Brady, B.H., 1999. Linking stress-dependent effective porosity and hydraulic
conductivity fields to RMR. International Journal of Rock Mechanics and Mining Sciences 36, 581-596.
Mavor, M.J., Vaughn, J.E., 1998. Increasing Coal Absolute Permeability in the San Juan Basin Fruitland
Formation. SPE Reservoir Evaluation & Engineering.
Mora, C.A., Wattattenbarger, R.A., 2009. Comparison of Computation Methods for CBM Performance.
Ouyang, Z., Elsworth, D., 1993. Evaluation of groundwater flow into mined panels. International Journal of
Rock Mechanics and Mining Sciences & Geomechanics Abstracts 30, 71-79.
Palmer, I., 2009. Permeability changes in coal: Analytical modeling. International Journal of Coal Geology 77,
119-126.
Palmer, I., Mansoori, J., 1998. How Permeability Depends on Stress and Pore Pressure in Coalbeds: A New
Model. SPE Reservoir Evaluation & Engineering.
Palmer, I.D., 2010. The Permeability Factor in Coalbed Methane Well Completions and Production, SPE
Western Regional Meeting. Society of Petroleum Engineers, Anaheim, California, USA.
Pekot, L.J., Reeves, S.R., 2002. Modeling the Effects of Matrix Shrinkage and Differential Swelling on
Coalbed Methane Recovery and Carbon Sequestration. U.S. Department of Energy.
Ranjbar, E., Hassanzadeh, H., 2011. Matrix–fracture transfer shape factor for modeling flow of a
compressible fluid in dual-porosity media. Advances in Water Resources 34, 627-639.
Reeves, S., Pekot, L., 2001. Advanced Reservoir Modeling In Desorption-Controlled Reservoirs, SPE Rocky
Mountain Petroleum Technology Conference. Copyright 2001, Society of Petroleum Engineers Inc.,
Keystone, Colorado.
Remner, D.J., Ertekin, T., Sung, W., King, G.R., 1986. A Parametric Study of the Effects of Coal Seam
Properties on Gas Drainage Efficiency. SPE Reservoir Engineering.
Ried, G.W., Towler, B.F., Harris, H.G., 1992. Simulation and Economics of Coalbed Methane Production in
the Powder River Basin, SPE Rocky Mountain Regional Meeting. 1992, Casper, Wyoming.
Roadifer, R.D., Moore, T.R., 2009. Coalbed Methane Pilots--Timing, Design and Analysis. SPE Reservoir
Evaluation & Engineering.
Robertson, E.P., 2005. Modeling Permeability in Coal Using Sorption-Induced Strain Data, SPE Annual
Technical Conference and Exhibition. Society of Petroleum Engineers, Dallas, Texas.
Sawyer, W.K., Zuber, M.D., Kuuskraat, V.A., Horner, D.H., 1987. Using reservoir simulation and field data
to define mechanisms controlling coalbed methane production, the 1987 Coalbed Methane
Symposium, Tuscaloosa.
Seidle, J.R., Huitt, L.G., 1995. Experimental Measurement of Coal Matrix Shrinkage Due to Gas Desorption
and Implications for Cleat Permeability Increases, International Meeting on Petroleum Engineering.
1995 Copyright 1995, Society of Petroleum Engineers, Inc., Beijing, China.
Shi, J.Q., Durucan, S., 2004. Drawdown Induced Changes in Permeability of Coalbeds: A New Interpretation
of the Reservoir Response to Primary Recovery. Transport in Porous Media 56, 1-16.
Shi, J.-Q., Durucan, S., 2008. Modelling of Mixed-Gas Adsorption and Diffusion in Coalbed Reservoirs, SPE
Unconventional Reservoirs Conference. Society of Petroleum Engineers, Keystone, Colorado, USA.
Siriwardane, H.J., Gondle, R.K., Smith, D.H., 2009. Shrinkage and swelling of coal induced by desorption and
sorption of fluids: Theoretical model and interpretation of a field project. International Journal of
Coal Geology 77, 188-202.
Thomas, L.K., Dixon, T.N., Pierson, R.G., 1983. Fractured Reservoir Simulation. Society of Petroleum
Engineers Journal, 42-54.
School of Mechanical and Chemical Engineering
Chapter 7 The University of Western Australia
7-34
Ueda, Y., Murata, S., Watanabe, Y., Funatsu, K., 1989. Investigation of the Shape Factor Used in the Dual-
Porosity Reservoir Simulator, SPE Asia-Pacific Conference. 1989 Copyright 1989, Society of
Petroleum Engineers, Inc., Sydney, Australia.
Warren, J.E., Root, P.J., 1963. The Behavior of Naturally Fractured Reservoirs.
Wilson, R.K., Aifantis, E.C., 1982. On the theory of consolidation with double porosity. International Journal
of Engineering Science 20, 1009-1035.
Wu, Y., Liu, J., Chen, Z., Elsworth, D., Pone, D., 2011. A dual poroelastic model for CO2-enhanced coalbed
methane recovery. International Journal of Coal Geology 86, 177-189.
Wu, Y., Liu, J., Elsworth, D., Miao, X., Mao, X., 2010. Development of anisotropic permeability during
coalbed methane production. Journal of Natural Gas Science and Engineering 2, 197-210.
Zhang H.B., Liu J., Elsworth D., 2008. How sorption-induced matrix deformation affects gas flow in coal
seams: A new FE model. International Journal of Rock Mechanics & Mining Sciences 45, 1226–1236
Ziarani, A.S., Aguilera, R., Clarkson, C.R., 2011. Investigating the effect of sorption time on coalbed methane
recovery through numerical simulation. Fuel 90, 2428-2444.
Zimmerman, R.W., Chen, G., Hadgu, T., Bodvarsson, G.S., 1993. A numerical dual-porosity model with
semianalytical treatment of fracture/matrix flow. Water Resour. Res. 29, 2127-2137.
School of Mechanical and Chemical Engineering
Chapter 8 The University of Western Australia
8-1
Chapter 8
Concluding Remarks
8.1 Main findings
This thesis has focused on the experimental and numerical study on geomechanics of coal-
gas interactions, particularly its effect on coal permeability evolution. This effort brings
together various aspects regarding the experimental measurements on coal properties to
different gas injections with different gas pressures and temperatures, directional
permeability model development, the roles of effective stress coefficient and sorption-
induced strain on permeability evolution under stress-controlled boundary conditions, and
the related permeability development, mechanism investigation on coal permeability
evolution and forecast of CBM production to various parameters with dual- porosity/dual-
permeability model. The major contributions and findings made in this study are
summarized below.
1. Chapter 2 presents the experimental measurement of adsorption isotherms, swelling,
permeability, and geomechanical properties for an Australian coal sample in N2, CH4
and CO2 at 35 ºC and 40 ºC. The results show that the gas adsorption is less at higher
temperature, and coal swelling strain with respect to adsorbed amount is close to the
same trend line irrespective of gas type and temperature. Permeability coefficients
decline with increased gas pressure, but it does not show a trend in relation to the gas
adsorption capacity for this coal. Elastic modulus of coal is found to be larger at higher
effective stress and lower temperatures.
2. In chapter 3, a directional permeability model was developed to define the evolution of
gas sorption-induced permeability anisotropy under in-situ stress conditions. This was
implemented into a fully coupled finite element model of coal deformation and gas
transport in a coal seam, and found that
The directional permeability of coal is determined by the mechanical boundary
conditions, the ratio of coal bulk modulus to coal matrix modulus, the initial
fracture porosity, and the magnitude of the sorption-induced coal swelling strain.
The boundary conditions control the magnitudes of total strains while the
School of Mechanical and Chemical Engineering
Chapter 8 The University of Western Australia
8-2
modulus reduction ratio partitions the effective strain (total strain minus the
swelling strain) between fracture and matrix.
For restraint conditions of uniaxial strain and for a constant volume reservoir,
changes in coal permeability are determined only by the gas pore pressure and the
swelling strain.
3. Chapter 4 presented the experimental observations of a series of gas flow-through
experiments all under the conditions of a constant pressure difference between
confining stress and pore pressure. Comparison between measured permeability data
and calibrated results demonstrates that the effective stress coefficient cannot be
assumed to be unity for both two coal samples and it could play an important role in the
evaluation of permeability change in adsorbing gasses, particularly with high gas flow
pressures.
4. A phenomenological permeability model has been developed in chapter 5 to explain
why coal permeability decreases even under the unconstrained conditions of variable
stress. Based on this permeability model, numerical results illustrate that coal
permeability profiles under the controlled stress conditions are mainly regulated by the
following factors: when the confining stress or swelling capacity is higher, the final
equilibrium coal permeability is much lower than the initial permeability; the
permeability increases more when the facture compressibility or effective stress
coefficient is higher; when the sensitivity factor is higher, the reduction in coal
permeability is more significant.
5. The performance of current coal permeability models was evaluated in Chapter 6 against
analytical solutions for the two extreme cases of either unconstrained swelling or
constrained However, these models have a boundary mismatch because model
boundary is constrained while experiment boundary is free displacement or
unconstrained. In this study, a matrix-fracture model is numerically investigated after
incorporating heterogeneous distributions of Young's modulus, Langmuir strain
constant in the vincity of the fracture. Comparison between homogeneous and
heterogeneous models has also been conducted. Major findings include:
Both homogeneous and heterogeneous models experience the swelling transition
from local swelling to macro swelling. At the initial stage of gas injection, matrix
swelling is localized within the vicinity of the fracture compartment. As the
School of Mechanical and Chemical Engineering
Chapter 8 The University of Western Australia
8-3
injection continues, the swelling zone is widening further into the matrix and
becomes macro-swelling.
Coal permeability experiences a rapid reduction at the early stage, a switch in
behavior from permeability reduction to recovery is observed, and coal
permeability finally recovers until it reaches the final equilibrium permeability. For
the homogeneous model, the final equilibrium permeability is always higher than
the original value, but the opposite was obtained for the heterogeneous model.
With the heterogeneous distributions of coal physical and swelling properties, this
numerical model matches with the experimental data reasonably well, which
demonstrates that heterogeneity of coal properties can be the answer to coal
permeability reduction under unconstrained conditions.
6. In Chapter 7, the dual-porosity/dual-permeability model was used to conduct a fully
coupled finite element model to quantify the impact of various parameters on the
production of coalbed methane (CBM). It couples coal deformation, gas flow and
transport in the matrix system, and gas flow in the fracture system. Based on the
modeling results, the following conclusions can be drawn:
The CBM production rate is controlled by interactions of multiple processes
triggered by the gas extraction itself. This unique feature of CBM reservoirs is
characterized through defining a number of coal properties such as permeability,
and coal-gas interactive properties such as desorption time constant, swelling and
desorption capacities, as a function of the gas extraction process.
The results of this study demonstrate important non-linear impacts of coal
property processes and coal-gas interactive property processes on CBM
productions. These impacts cannot be recovered by previous studies where
mechanical influences are not rigorously coupled with the complex interactions
between fractures and matrixes.
The results of this study also demonstrate the importance of appropriate coal
property models and coal-gas interactive property models for the successful
forecast of CBM productions because these properties are very sensitive to the gas
extraction process.
These conclusions above suggest that the optimal in-situ control of coal properties and
coal-gas interactive properties could be achieved through designing a better gas
extraction process.
School of Mechanical and Chemical Engineering
Chapter 8 The University of Western Australia
8-4
8.2 Recommendations for future work
In this thesis, the comprehensive study of the geomechanics of coal-gas interactions on the
evolution of coal properties, coal permeability in particular, has been conducted, but
further investigations are important and still need research efforts in the future as outlined
below:
1. Whether adsorption causes coal softening is not clear, since the Young′s modulii with
Helium are not higher than those measured when adsorbing gasses. More work needs to
be done to further elucidate the impact of adsorption on geomechanical properties.
2. Our experimental results show the cleat compressibilities at 40 ºC are smaller than those
in 35 ºC. This may be because of the lower swelling strains at higher temperature leading
to less porosity decrease. However, since the cleat aperture may be altered by increasing
the temperature, the conclusion of lower compressibility at higher temperature will
require further investigation.
3. Due to the nature of low permeability, coal seams usually need to be hydraulically
fractured before coalbed methane recovery process begins. The fracture permeability is
not constant but changes with in situ conditions during production, so it is a challenge
to properly include and simulate the property permeability change in geomechanical
simulation.
4. Although it has been reported that the gas flow in the cleat system may be of the non-
Darcy nature, little has been known on how this non-Darcy flow affects the coalbed
methane extraction. Therefore, the non-Darcy effect is an interesting topic needs to
look at in the future.
5. Most studies on coal-gas interactions were focused on single gas component. How
multiple gasses interact with coal matrixes and fractures, especially how gas components
affect the geomechanical deformation, are not fully understood. Further research on this
issue could be important.