Sensitivity Study of Coalbed Methane
Production With Reservoir and
Geomechanic Coupling Simulation
F. GU, R.J. CHALATURNYK
University of Alberta
Abstract
Permeability of coal seams is one of the key factors for the
success of coalbed methane (CBM) developments. It is domi-
nated by cleat permeability in coal, which is very sensitive to
the change of effective stresses. The coal matrix shrinkage due
to methane production also influences cleat permeability. Using
an explicit-coupling simulation method, which simultaneously
simulates multiphase fluid flow and coal deformation, and a
coupling permeability model, which considers the effects of
the effective stress change and coal matrix shrinkage on cleat
permeability, the sensitivity of CBM production to ten engi-
neering, geologic, and coal intrinsic parameters such as cleat
permeability, cleat spacing, well control area, depth, and methane
content, etc., were studied in this paper. These parameters are
stress and matrix shrinkage related parameters or have signifi-
cantly influences on CBM production identified from previous
studies. The production rate and final gas recovery from con-
ventional simulations and coupling simulations are also com-
pared. Of the parameters studied, permeability, cleat spacing, and
in situ stresses were found to be the most sensitive parameters
that influence CBM production. Medium sensitivity was found
for the coefficient of matrix shrinkage, the Langmuir volume,
pressure gradient, and well control area, while the least sensitive
parameters included Poissons ratio, Youngs modulus, and the
Langmuir pressure.
Introduction
With reserves of 84 ~ 262 trillion m3 (2,980 ~ 9,260 trillion
ft3) all over the world(1), coalbed methane (CBM) has come to
represent a real gas supply to meet current and future natural gas
demands. In the United States, CBM accounted for 10% of dry gas
reserves and 8% of dry gas production in 2003(2). The worldwide
development of CBM is also accelerating in many other countries
such as China, Canada, and Australia.
The success of CBM developments depends on many factors,
but specific properties of a coal seam remain the fundamental
controlling factor. Many people have investigated the effects of
coal seam properties on CBM production and recovery(3-6). The
studies of Sawyer et al. indicated that cleat (fracture) permeability
and relative permeability, not gas diffusion, control long-term
productivity, and that optimum well spacing also depends on
cleat permeability(3). Reid et al.s results showed that permea-
bility, initial desorption pressure, and drainage area are the most
important reservoir parameters for CBM production(4). Young et al.
pointed out that permeability, well spacing, and the degree of coal
saturation have the greatest impacts on the long-term performance
of CBM wells(5). A completed parametric study by Roadifer
PEER REVIEWED PAPER (REVIEW AND PUBLICATION PROCESS CAN BE FOUND ON OUR WEB SITE)
October 2005, Volume 44, No. 10
et al. illustrated that for coal-only reservoirs (without adjacent sand
layers), the five parameters having the most impact on the peak gas
rate are, in order of highest to lowest, permeability, free gas satu-
ration, degree of saturation of the coal, damage skin factor, and
thickness(6).
The results of the above-mentioned studies clearly indicate that
cleat permeability is likely the most important factor for CBM
production. However, in all these investigations, permeability was
considered as a constant. Actually, it is not only a variable, but also
changes drastically during CBM production due to the alteration
of in situ conditions, such as pressure changes and the production
of methane.
The decrease of pore fluid pressures during CBM production
will result in an increase of in situ effective stresses that act on
the cleats and cause them to be compressed or decrease in aper-
ture. This effect will lead to a permeability decrease. The strong
stress dependency of coal permeability has been recognized for
a long time. Using N2 and CO2, Patching found that coal perme-
ability decreased approximately by four orders of magnitude while
the confining stress increased from 0.07 to 20.68 MPa (10 to 3,000
psi)(7). The investigations of Somerton et al. showed the permea-
bility of fractured coal to methane decreased by more than two or-
ders in the stress range of 0.34 to 13.79 MPa (50 to 2,000 psi)(8).
Reznik et al. measured coal permeability to nitrogen and found
it decreased approximately two orders when the confining pres-
sure increased from 0.689 to 4.83 MPa (100 to 700 psi)(9). The re-
sults of Rose and Foh indicated that coal permeability to water also
decreased as much as two orders of magnitude over the pressure
range of normal dewatering and production cycles(10). Enever and
Hening illustrated that the logarithm of coal permeability linearly
decreases with the increase of effective stresses(11). Seidle et al.(12)
theoretically deduced the logarithm of permeability ratio linearly
decreases with the change of horizontal stresses. Schwerer and
Pavones study indicated that the pressure-dependant permeability
phenomenon significantly reduced the production of water and gas
compared with that expected on the basis of a constant permea-
bility equal to the value at initial coal seam conditions(13).
Another important change of in situ conditions is the extraction
of methane. The coal matrix will shrink due to the desorption of
methane. The matrix shrinkage produces a relaxation of in situ
stresses resulting in the increase in apertures and permeability of
cleats. Harpalani and Zhaos studies showed that under pressures
higher than the desorption pressure of methane (without methane
desorption) coal permeability decreased with decreasing gas pres-
sures. However under pressures lower than the desorption pres-
sures (with methane desorption) coal permeability to methane
drastically increased with decreasing gas pressures(14). In contrast,
the permeability to helium (almost non-adsorptive gas) decreases
continuously with decreasing gas pressure. This clearly showed
that the shrinkage of coal due to methane desorption caused the
23
permeability increase of coal. Harpalani and Schraufnagels
results also indicated that coal permeability to methane increases
with decreasing gas pressures, in spite of the increase of effective
stresses(15). Based on the tested shrinkage coefficient of San Juan
coal, Seidle and Huitt estimated that if the coalbed fluid pressure
was decreased from 10.342 MPa (1,500 psi) to 0.689 MPa (100
psi), the porosity would increase 20% and the permeability would
increase 70% due to the shrinkage of coal(16). From field well tests
of three wells, Mavor and Vaughn illustrated that the absolute per-
meability increased as much as 2.7 to 7 times comparing with their
original values after produced 3 4 years(17).
Therefore, coal permeability decreases due to pressure declines
while increases due to matrix shrinkage during CBM production.
These opposing effects will partially offset each other. However,
the degree to which each factor influences the magnitude of per-
meability change varies with production time and location within
a coal seam. Consequently, a proper treatment of both mechanisms
is necessary to improve the reliability of simulation predictions of
CBM production.
There are two types of methods that can be used to include
the dynamic change of permeability in the simulation of CBM
production. One method is to use analytical models to calculate
the permeability changes as a function of fluid pressure changes
and apply this relationship directly in reservoir simulations. Many
analytical models have been developed, however, a discussion
of these models is beyond the scope of this paper. Gu and Cha-
laturnyk(18) provide a discussion of many of these models and, in
general, the models put forward by Sawyer and Paul(19), Palmer
and Mansoori(20), and Shi and Durucan(21) represent the most suit-
able analytical models since they capture both the decline of pore
pressures and the coal matrix shrinkage due to gas desorption.
In these models, a monotonic relation between in situ stresses
FIGURE 1: Fracture system with matchstick matrix blocks.
FIGURE 2: Physical models for fluid flow simulation and
geomechanical simulation.
24
and pore pressures are obtained and used based on their
assumptions. The study by Gu and Chalaturnyk showed that during
CBM production, in situ stress changes depend not only on fluid
pressure changes but also on spatial position within a coal seam.
Using 3D simulations, they were able to show that no unique
(monotonic) relationship exists between in situ stresses and pore
pressures(18). This result suggests caution is required in applying
these analytical models in CBM production simulations.
The second method to model the dynamic change of permeability
is to utilize geomechanical and reservoir coupling simulations, in
which the geomechanical simulation is implemented for general-
ized deformation and stress change predictions, while multiphase
flow is simulated with an appropriate reservoir simulator(22, 23).
The following section describes the sensitivity study con-
ducted to examine the influence of reservoir parameters related to
in situ stresses and matrix shrinkage. A complete sensitivity study
of CBM production, such as one done by Roadifer et al.(6), is not
the goal of this paper, but the factors that most significantly affect
CBM production identified from previous studies will be reviewed
in this paper.
Description of Coupling Simulation
Coal seams are idealized as a matchstick fracture system(12), as
shown in Figure 1. Coal blocks are intersected by two orthogonal
and homogenous fracture or cleat sets representing face cleats and
butt cleats. The initial apertures of two cleat sets are assumed to be
equal in this study (i.e., permeability is initially isotropic), though
Gu and Chalaturnyks coupling permeability model can handle the
anisotropy of permeability(22).
The physical models for both multiphase flow simulations
and geomechanical simulations are illustrated in Figure 2. In
both simulations, the same variable radial grids are used, i.e.,
the grid sizes are smaller near the wellbores and larger near the
boundary with grid dimensions that sequentially increase in size
at a ratio of 1.2:1. For the fluid flow simulation, the grids are 30
(radial) 9 (perimeter) 1 (vertical) and the grid sizes are equal
in the perimeter direction. For the geomechanical simulation, the
grids are 30 (radial) 22 (vertical), of which two grids are distrib-
uted to coal seams in the vertical direction, and the grid sizes in
the overburden are also variable, i.e., the smaller sizes are adja-
cent to the coalbed and the larger sizes are near the upper boundary
with grid dimensions that sequentially increase in size at a ratio
of 1.1:1.
An explicitly sequential coupling simulation methodology is
used for the reservoir and geomechanical coupling simulation in
this study. Fluid flow is calculated with GEM(24) and geome-
chanical deformation is calculated with FLAC(25). The former is
a multidimensional, multiphase, isothermal, compositional simu-
lator and has the capability of coalbed methane simulation. The
latter is a widely used geomechanical code that is designed for
rock and soil mechanics analyses. It can simulate thermo-me-
chanical, hydro-mechanical, and thermo-hydro-mechanical in-
teractions. Additional codes are necessary to perform selective
FIGURE 3: Procedure of reservoir and geomechanic coupling
simulation.
Journal of Canadian Petroleum Technology
computations and exchange data between the two simulators. The
whole coupling simulation job is managed by AutoMate and can
execute automatically. The procedure for the coupled simulations
is shown in Figure 3, where the calculations of an arbitrary time
step (time step i) are indicated with a dotted outline region. The
sub-calculation steps in one time step include:
1. Calculate the pressures (P), absorbed gas volumes of (VD),
water saturation (SW), and well production rates of gas and
water (Qg, QW) etc. using GEM;
2. Deal with the results of step (1) and prepare the pressures (P)
and absorbed gas volumes (VD) in the format that can be used
by FLAC with a self-programmed code;
3. Calculate stresses (m) and linear strains (l) with the new pres-
sures and absorbed gas volumes from step (2) with FLAC;
and,
4. Calculate cleat permeability (kf) with Equation (1) showed
in the following based on the linear strains (l) of Step 3, and
prepare the permeability data in the format that can be used
by GEM with a self-programmed code.
In the coupling simulation, the change of cleat permeability
during production is predicted with Gu and Chalaturnyks coupling
permeability model(22):
kf a 3
= 1 + l
kf 0 b
.................................................................................. (1)
Note that compressive strains are positive. In this model, the repre-
sentation of a unit consisting of a coal matrix block plus a cleat in
the bedding plane is assumed unchanged during the production.
The change of linear strain is calculated by FLAC and can be
expressed as:
l = lS + lP + lD + lT
........................................................... .(2)

TABLE 1: The parameters of coal seam. In this study, the production of methane is considered as an iso-
thermal process, thus the change of thermal strains is zero. Be-
Seam thickness (m) 10 cause cleat volume is much smaller compared to matrix volume,
Coal density (kg/m3) 1,542 the influence of cleat volume on methane storage is assumed neg-
Well radius (m) 0.1
ligible and so, cleat porosity is fixed in this study.
Temperature of surface constant layer ( C) 8
Temperature grad ( C/100m) 2.5
Water saturation in fracture (fraction) 1
Matrix porosity (fraction) 0.005 Effects of Coal Deformation
Matrix permeability (mD) 0.001
Water viscosity (cp) 0.644 Prior to exploring the sensitivity study, it is instructive to
Water density (kg/m3) 990 investigate how the deformation or strain change of a coal mass
Rock compressibility (1/kPa) 1.45E-07
Water compressibility (1/kPa) 5.80E-07 TABLE 3: Production constraints.
Reference pressure of compressibility (kPa) 8,500
Methane sorption time (day) 100 Water Production Constraint
Time (day) 0 ~ 18
Maximum water rate (m3/day) 10
TABLE 2: The parameters of overburden. Minimum BHP (kPa) 101.325

Density (kg/m3) 2,300 Gas Production Constraint

Youngs modulus (MPa) 1,000 Time (day) 19 ~ 7,300
Poissons ratio 0.3 Minimum BHP (kPa) 275

TABLE 4: Parameters of sensitivity study.

No Items Unit Lower Value Base Value Upper Value

Permeability mD 0.2 4 80
1 Cleat Width micron 3.36 10 26.8
Porosity fraction 0.000363 0.001 0.00268
2 Cleat space m 0.005 0.02 0.1
Coefficient of matrix
3 g/ml 1.0 10-4 4.0 10-4 7.0 10-4
shrinkage
Area acres 50 175 300
4 Well control
Radius m 253.8 474.8 621.6
psi/ft 0.33 0.43 0.53
Gradient
5 Pressure kPa/m 7.465 9.727 11.989
KPa 6,718.5 8,754.3 10,790.1
6 Depth m 300 900 1,500
7 Youngs modulus of coal psi 145,000 493,000 725,000
MPa 999.7 3399.2 4998.7
8 Poissons ratio of coal 0.22 0.32 0.42
ft3/ton 300 800 1300
9 Langmuir volume
m3/ton 8.495 22.653 36.812
psi 200 600 1,000
Pressure
kPa 1,379.0 4,136.9 6,894.8
10 Langmuir
Pressure
1/kPa 7.252 10-4 2.417 10-4 1.450 10-4
reciprocal

October 2005, Volume 44, No. 10 25

 FIGURE 4: The effects of coal deformation on well rate.
FIGURE 5: The effects of coal deformation on cumulative
production.
can influence CBM production. The coalbed is assumed to have
isotropic geomechanical properties. The data of the well are listed
in Tables 1, 2, and 3 and in the column showing base value in
Table 4.
In order to study the influence of anisotropic deformations of
coal on CBM production, two average methods are applied for the
linear strain in Equation (1). One method is to use one third of the
volumetric strain as the linear strain and the other is to use the av-
erage horizontal linear strain (i.e., averaging the linear strains in
the bedding plane or x-y plane in Figure 1) as the linear strain.
Figures 4 and 5 show the effects of coal deformation on CBM
production. For comparison, the case of zero linear strain, i.e.,
corresponding to constant permeability case, is also plotted in
these figures. The results indicate that when the coal deformation
is considered, the production rates and cumulative gas recovery are
higher than that for no deformation (zero strain). This occurs be-
cause during production, the coal matrix shrinkage due to methane
desorption dominates the deformation and leads to increasing cleat
permeability. Figures 4 and 5 also show that the production rates
and methane recovery simulated with the linear strain equal to
one third of the volumetric strain are much higher than that with
the linear strain equal to the average horizontal linear strain. The
reason this occurs is that even though coal is assumed to have iso-
tropic mechanical properties, the strain changes are anisotropic,
i.e., the linear strain change in the vertical direction, which does
not actually contribute to the horizontal permeability for fluid flow,
are larger than the linear strain changes in the horizontal plane.
In the following sensitivity study, the average horizontal linear
strains are utilized for all simulations.
26
FIGURE 6: The effects of permeability on well rate.
FIGURE 7: The effects of permeability on cumulative production.
Sensitivity Study
The investigations of this study focus on the factors that are re-
lated to in situ stresses and the coal matrix shrinkage due to de-
sorption with an incorporated explicit-coupling simulation. Those
factors having been recognized as the most significant factors that
influence CBM production from previous studies will also be in-
vestigated here. The well data are listed in Tables 1 and 2 and the
well constraints for production are shown in Table 3. The param-
eters and their values for the sensitivity study are shown in Table
4. The ranges of these parameters are from several sources, mainly
from the studies of Levine(26), Mavor et al.(27), and Roaldifer et
al(6). The base values in Table 4 are used in all coupling simula-
tions except for the parameter that is investigated in each sensi-
tivity study.
Cleat Permeability
The changes of cleat permeability and corresponding cleat
widths and porosity are shown in Table 4. It should be noted that
with the fixed cleat spacing, i.e., 0.02 m, the lower, base, and upper
values of porosity are calculated by the following relationship(28):
2b
f =
a .................................................................................................. (3)
The effects of cleat permeability on CBM production are shown
in Figures 6 and 7. As expected, the results show that cleat perme-
ability has a significant influence on CBM production. The higher
the cleat permeability is, the higher the maximum rate and cumu-
lative recovery are.
Journal of Canadian Petroleum Technology
 FIGURE 8: The effects of permeability on production from
conventional simulation.
FIGURE 9: The effects of cleat space on well rate.
FIGURE 10: The effects of cleat space on cumulative production.
The results from conventional simulations (no coupling or
constant permeability) are shown in Figure 8. Comparing these
results with those shown in Figures 6 and 7 indicates that the
maximum gas rate and final gas recovery from coupling simu-
lations are higher than those from the simulations with constant
permeability for the same cleat permeability. This means cleat
permeability is very sensitive to the shrinkage of coal.
Cleat Space
In the study on the influence of cleat space on CBM production,
porosity was adjusted based on Equation (3) in order to make cleat
permeability equal to the base value. The simulation results are
plotted in Figures 9 and 10. The results show that with the increase
October 2005, Volume 44, No. 10
FIGURE 11: The effects of cleat space on production from
conventional simulation.
FIGURE 12: The effects of coefficient of matrix shrinkage on well
rate.
of cleat space, the maximum gas rate and final gas recovery sig-
nificantly increase but the time to reach the maximum gas rate de-
creases. The primary reason for this effect is that as the cleat space
becomes larger for the same linear strain change, the permeability
increase is higher according to Equation (1).
When comparing these coupling simulation results with that
from conventional simulations plotted in Figure 11, it shows
that the maximum gas and final gas recovery from coupling
simulations are, again, much higher than that from conventional
simulation. This comparison shows again that the influence of coal
matrix shrinkage due to methane exaction on CBM production is
significant.
Coefficient of Matrix Shrinkage
The coefficient of coal matrix shrinkage is defined according to
the assumption that the volumetric strain of coal due to methane
desorption is linearly related with absorbed methane(22), that is:
vD = D VD
...................................................................................... (4)
The coefficient of coal matrix shrinkage, `D, depends on coal type
and is determined experimentally.
The results from coupling simulations exploring the influence
of the coefficient of coal matrix shrinkage due to methane desorp-
tion are shown in Figures 12 and 13. These results suggest that for
the seam with a higher coefficient of matrix shrinkage, the max-
imum rates and final gas recovery are higher as compared with a
seam having a lower coal shrinkage coefficient. This is expected
since a higher coefficient of matrix shrinkage will cause a larger
shrinkage, i.e., a larger compressive linear strain, thus a higher
increase in permeability from Equation (1).
27
 FIGURE 13: The effects of coefficient of matrix shrinkage on
cumulative production.
FIGURE 14: The effects of well control area on well rate.
FIGURE 15: The effects of well control area on cumulative
production.
Well Control Area
The effects of well control area on CBM production are indicated
in Figures 14 and 15. These results illustrate that for a well with
a larger control area, the maximum gas rate is lower and the time
to achieve the maximum rate is longer. However, the final gas re-
covery is higher than in wells with a smaller control area due to
more initial gas in place.
The simulation results from conventional simulations (no cou-
pling) are shown in Figure 16. The comparison between the results
from coupling simulations and those from conventional simu-
lations shows that the gas rate and final gas recovery from cou-
pling simulations are also higher than those from conventional
simulations.
28
FIGURE 16: The effects of well control area on production from
conventional simulation.
FIGURE 17: The effects of pressure gradient on well rate.
FIGURE 18: The effects of pressure gradient on cumulative
production.
Pressure Gradient
The pressure gradient of a reservoir is defined as the quotient of
the initial reservoir pressure divided by the depth of the reservoir.
The corresponding reservoir pressures shown in Table 4 are
obtained by multiplying the pressure gradients with the fixed res-
ervoir depth, i.e., 900 m in this study.
Simulation results examining the sensitivity of pressure gradient
on CBM production are illustrated in Figures 17 and 18. These
results imply that the wells with higher pressure gradients have
higher maximum gas rates, but the time to achieve the maximum
rate does not change notably. The final gas recovery from wells
with a higher pressure gradient is also higher than from those with
lower pressure gradients.
Journal of Canadian Petroleum Technology
 FIGURE 19: The effects of pressure gradient on production from
conventional simulation.
FIGURE 20: The effects of depth on well rate.
FIGURE 21: The effects of depth on cumulative production.
A similar conclusion is also reached based on the results of
conventional simulations shown in Figure 19. The comparison of
Figure 19 with Figures 17 and 18 clearly shows that the maximum
rates and final recovery are all higher, but the time to reach the
maximum gas rate is lower for coupling simulations than for con-
ventional simulations.
Seam Depth
In this study, since the pressure gradient is fixed, i.e., 9.727 kPa/
m, the initial reservoir pressures are obtained by multiplying this
pressure gradient with the reservoir depths. Note that the initial in
situ total stresses also need to be re-calculated due to the changes
of depth based on the fixed rock density of overburden and coal,
October 2005, Volume 44, No. 10
FIGURE 22: The effects of Youngs modulus of coal on well rate.
FIGURE 23: The effects of Youngs modulus of coal on cumulative
production.
and on the assumption that the ratio of the horizontal stress to
vertical stress is one (1).
The coupling simulation results for the influence of seam depth
are plotted in Figures 20 and 21. The results indicate that for an in-
creasing seam depth, which reflects the magnitude of initial in situ
stresses, the gas rate and gas recovery of deeper wells are much
higher than that of shallower wells. This is a misleading conclu-
sion, however, since the sensitivity study has assumed that all other
parameters (i.e., cleat permeability, cleat space, and pressure gra-
dient) in the simulation maintain baseline values. In reality, this
would not be true, especially for cleat permeability, which de-
creases with an increase of seam depth.
Youngs Modulus of Coal
The coupling simulation results for the influence of Youngs
modulus of coal are illustrated in Figures 22 and 23. These re-
sults imply that maximum gas rates and gas recovery from a coal
seam with a smaller modulus are higher than those with a larger
modulus. But the time to reach the maximum gas rate is almost the
same. This occurs because as the pore pressures decrease during
production, the effective stresses increase, and coal with a smaller
Youngs modulus will compress more than a coal with a larger
Youngs modulus as a result of the increase in effective stresses.
Increased compression creates more space for the cleats based on
the assumption that the total width of a coal matrix block plus the
adjacent cleat is constant in the coupling permeability model.
However, this conclusion is subject to the condition that the
cleats are strong enough and deform elastically. The influence
of the rock discontinuity between the matrix and cleat (fracture)
has not been included in this study. Further study of this issue is
required.
29

FIGURE 24: The effects of Poissons ratio of coal on well rate.
FIGURE 25: The effects of Poissons ratio of coal on cumulative
production.
FIGURE 26: The effects of Langmuir volume on well rate.
Poissons Ratio of Coal
The influences of Poissons ratio on CBM production are illus-
trated in Figures 24 and 25. The results suggest that the maximum
gas rates and gas recovery from wells with coal having a higher
Poissons ratio are lower than from wells with coal having a lower
Poissons ratio. The difference in the time to reach the maximum
gas rates is insignificant. The reason for this effect follows the
same volume change or compression argument presented above
for Youngs modulus.
Langmuir Volume
The effects of Langmuir volume on CBM production are shown
in Figures 26 and 27. These results indicate that with an increase
of the Langmuir volume in a coal seam, the maximum gas rate and
30
FIGURE 27: The effects of Langmuir volume on cumulative
production.
FIGURE 28: The effects of Langmuir volume from conventional
simulation.
final gas recovery all increase. The time to reach the maximum
gas rate also slightly increases with an increase in the Langmuir
volume. This is due to an increase in situ gas volume with the in-
crease of Langmuir volume, while the pressure of the coal seam is
fixed. This results in a larger shrinkage of the coal matrix during
production, and thus a higher permeability increase.
The comparison of the coupling simulation results with con-
ventional simulation results illustrated in Figure 28 shows that the
maximum gas rate and final gas recovery of coupling simulations
are higher. The time to achieve the maximum gas rate is shorter
for coupling simulations compared with that of conventional
simulations.
Langmuir Pressure
The coupled simulation results to study the influence of Lang-
muir pressure on CBM production are illustrated in Figures 29 and
30. These results suggest that with increasing the Langmuir pres-
sure, the maximum gas rate and final gas recovery increase. How-
ever, once the Langmuir pressure is greater than about half of the
coal seam pressure, the influence is negligible.
For comparison, the results from conventional simulations are
shown in Figure 31. The results show that the effects of Lang-
muir pressure reveal a similar trend as that in the coupling sim-
ulations. For the same Langmuir pressure, the gas rate and final
gas recovery from coupling simulations are higher than those from
conventional simulations. Similarly, matrix shrinkage is primarily
responsible for this effect.
Conclusion
From the simulation results and previous discussions, the
following conclusions are obtained:
Journal of Canadian Petroleum Technology
 FIGURE 31: The effects of Langmuir pressure from conventional
FIGURE 29: The effects of Langmuir pressure on well rate. simulation.

kf0 = permeability of fracture (cleat) at the initial state

`D = matrix shrinkage coefficient of coal matrix due to change
of absorbed gas
qf = fracture (cleat) porosity
6VD = change of absorbed methane volume, 6VD = VD0 VD
6l = change of linear strain
6lD = change of linear strain due to gas desorption
6lS = change of linear strain due to change of total stress
6lP = change of linear strain due to change of pore pressure
6lT = change of linear strain due to change of temperature
6VD = change of volumetric strain due to desorption/adsorption
of gas

REFERENCES
FIGURE 30: The effects of Langmuir pressure on cumulative
production.
1. Caution should be exercised in estimating permeability
changes from volumetric strains in coupling simulations. If
the anisotropic deformation of coal is not treated explicitly
and linear strains are simply computed as one third of the
volumetric strain, the simulated CBM production rates and
gas recovery may be overly optimistic;
2. Permeability, cleat spacing, and seam depth (corresponding
to in situ stresses) are the most sensitive parameters that
influence CBM production from coupling simulations;
3. The Langmuir volume (corresponding to initial gas content),
pressure gradient (related to initial gas content and initial
stresses), and well control area are the second most sensitive
parameters;
4. Coefficient of matrix shrinkage, Poissons ratio, Youngs
modulus, and the Langmuir pressure are the relatively least
sensitive parameters influencing CBM production; and,
5. Due to the consideration of matrix shrinkage of coal, the
maximum gas rate and final gas recovery from coupling
simulations are higher than those from conventional
simulations (constant permeability) when matrix shrinkage
dominates the deformation of the coal.
Acknowledgements
The first author would like to acknowledge the financial support
for this study provided by the Alberta Ingenuity Fund.
NOMENCLATURE
a = width of coal matrix block
b = aperture of fracture (cleat)
kf = permeability of fracture (cleat)
kf0 = permeability of fracture (cleat) at a reference condition
October 2005, Volume 44, No. 10
1. DAVIDSON, R.M., SLOSS, L.L., and CLARKE L.B., Coalbed
Methane Extraction; IEA Coal Research, London, January 1995.
2. Energy Information Administration of U.S. DOE, U.S. Crude Oil,
Natural Gas, and Natural Gas Liquids Reserves: 2003 Annual Report;
Washington DC, November 2004.
3. SAWYER, W.K., ZUBER, M.D., and KUUSKRAA, V.A., Using
Reservoir Simulation and Field Data to Define Mechanisms
Controlling Coalbed Methane Production; paper 8763, presented at
the Coalbed Methane Symposium, Tuscaloosa, AL, November 16
19, 1987.
4. REID, G.W., TOWLER, B.F., and HARRIS, H.G., Simulation and
Economics of Coalbed Methane Production in Power River Basin;
paper SPE 24360, presented at the SPE Rocky Mountain Regional
Meeting, Casper, WY, May 18 21, 1992.
5. YOUNG, G.B.C., PAUL, G.W., MCELHINEY, J.E., and MCBANE,
R.A., A Parametric Analysis of Fruitland Coal Methane Reservoir
Productivity; paper SPE 24903, presented at the 67th Annual
Technical Conference and Exhibition of the Society of Petroleum
Engineers, Washington DC, October 4 7, 1992.
6. ROADIFER, R.D., MOORE, T.R., RATERMAN, K.T., FARNAN,
R.A., and CRABTREE, B.J., Coalbed Methane Parametric Study:
Whats Really Important to Production and When?; paper SPE 84425,
presented at the SPE Annual Technical Conference and Exhibition,
Denver, CO, October 5 8, 2003.
7. PATCHING, T.H., Variations in Permeability of Coal; proceedings of
the Rock Mechanics Symposium, University of Toronto, ON, January
15 16, 1965.
8. SOMERTON, W.H., SOYLEMZOGLU, I.M., and DUDLEY, R.C.,
Effect of Stress on Permeability of Coal; International Journal of
Rock Mechanics and Mining Science & Geomechics Abstracts, Vol.
12, pp. 129-145, 1975.
9. REZNIK, A.A., LIEN, C.L., and FULTON, P.F., Permeability
Characteristics of Coal; proceedings of the 4th Underground Coal
Conversion Symposium, Steamboat Springs, CO, July 17 20, 1978.
10. ROSE, R.E. and FOH, S.E., Liquid Permeability of Coal as a Func-
tion of Net Stress; paper SPE/DOE/GRI 12856, presented at the
SPE/DOE/GRI Unconventional Gas Recovery Symposium, Pitts-
burgh, PA, May 13 15, 1984.
11. ENEVER, J.R.E. and HENING A., The Relationship Between
Permeability and Effective Stress for Australian Coals and its
Implications With Respect to Coalbed Methane Exploration and Res-
ervoir Modelling; paper 9722, presented at the International Coalbed
Methane Symposium, Tuscaloosa, AL, May 12 16, 1997.
31
 12. SEIDLE, J.P., JEANSONNE, M.W., and ERICKSON, D.J., Appli-
cation of Matchstick Geometry to Stress Dependent Permeability in Authors Biographies
Coals; paper SPE 24361, presented at the SPE Rocky Mountain Re-
gional Meeting, Casper, WY, May 18 21, 1992. Fagang Gu is a Ph.D. student at the Uni-
13. SCHWERER, F.C. and PAVONE, A.M., Effect of Pressure-Dependent versity of Alberta. Previously, he was an as-
Permeability on Well-Test Analyses and Long-Term Production of sociate professor at Southwest Petroleum
Methane From Coal Seams; paper SPE/DOE/GRI 12857, presented
Institute (SWPI) of China. His research
at the SPE/DOE/GRI Unconventional Gas Recovery Symposium,
Pittsburgh, PA, May 13 15, 1984. interests consist of CBM and ECBM re-
14. HARPALANI, S. and ZHAO, X., An Investigation of the Effect of covery, reservoir engineering, stimulation
Gas Desorption on Coal Permeability; paper 8923, presented at the and production optimization, etc. His ex-
Coalbed Methane Symposium, Tuscaloosa, AL, April 17 20, 1989. periences include new technology and new
15. HARPALANI, S. and SCHRAUFNAGEL, R.A., Shrinkage of Coal model study, engineering simulator devel-
Matrix With Release of Gas and its Impact on Permeability of Coal; opment, and treatment design and field
Fuel, Vol. 69, pp. 551-556, 1990. consultation. He holds B.Sc. and M.Sc. de-
16. SEIDLE, J.P. and HUITT, L.G., Experimental Measurement of grees in petroleum engineering. He is a member of SPE and the
Coal Matrix Shrinkage Due to Gas Desorption and Implications for
Petroleum Society.
Cleat Permeability Increases; paper SPE 30010, presented at the
International Meeting on Petroleum Engineering, Beijing, PR China,
November 14 17, 1995. Rick Chalaturnyk is an associate pro-
17. MAVOR, M.J. and VAUGHN, J.E., Increasing Coal Absolute Perme- fessor of geotechnical engineering at the
ability in the San Juan Basin Fruitland Formation; SPE Evaluation & University of Alberta, having obtained his
Engineering, pp. 201-206, June 1998. Ph.D. in geotechnical engineering from
18. GU, F. and CHALATURNYK, R.J., Numerical Simulation of Stress the University of Alberta. Before joining
and Strain due to Gas Sorption/Desorption and Their Effects on In the University, he was involved in petro-
Situ Permeability of Coalbeds; paper CIPC 2005-058, the Petroleum leum geomechanics research in SAGD,
Societys Canadian International Petroleum Conference, Calgary,
heavy oil cold production with the Centre
AB, June 7 9, 2005.
19. SAWYER, W.K. and PAUL, G.W., Development and Application of for Engineering Research, and co-founded
a 3D Coalbed Simulator; paper the Petroleum Society/SPE 90-119, a downhole instrumentation and moni-
The International Technical Meeting, Calgary, AB, June 10 13, toring company. He has been involved in a
1990. wide range of research activities ranging from mine tailings man-
20. PALMER, I. and MANSOORI, J., How Permeability Depends on agement technologies, geological storage of greenhouse gases,
Stress and Pore Pressure in Coalbeds: A New Model; SPE Evaluation gas-over-bitumen issues pertaining to SAGD development, and
& Engineering, pp. 539-544, December 1998. coalbed methane geomechanics. He is a registered professional en-
21. SHI, J.Q. and DURUCAN, S., Changes in Permeability of Coalbeds gineer in Alberta, and is a member of SPE, AGU, SRA, CGS, and
During Primary RecoveryPart 1: Model Gormulation and Snalysis;
the Petroleum Society.
paper 0341, presented at the International Coalbed Methane Sympo-
sium, Tuscaloosa, AL, May 5 9, 2003.
22. GU, F. and CHALATURNYK, R.J., Analysis of Coalbed Methane
Production by Reservoir and Geomechanical Coupling; paper 2003-
061, presented at the Petroleum Societys Canadian International
Petroleum Conference, Calgary, AB, June 10 12, 2003.
23. ZHAO, Y., HU, Y., ZHAO, B., and YANG, D., Nonlinear Coupled
Mathematical Model for Solid Deformation and Gas Seepage in
Fractured Media; Transport in Porous Media, Vol. 55, pp. 119-136,
2004.
24. CMG, Users Guide GEM; Computer Modelling Group Ltd., Calgary,
AB, 2002.
25. Itasca, FLAC (Fast Lagrangian Analysis of Continua) Users Guide;
Itasca Consulting Group, Inc., Minneapolis, MN, 2002.
26. LEVINE, J.R., Model Study of the Influence of Matrix Shrinkage
on Absolute Permeability of Coal Bed Reservoirs; Coalbed Methane
and Coal Geology (Geological Society Special Publication No. 109),
The Geological Society, London, 1996.
27. MAVOR, M.J., OWEN, L.B., and PRATT, T.J., Measurement and
Evaluation of Coal Isotherm Data; paper SPE 20728, presented at the
65th Annual Technical Conference and Exhibition, New Orleans, LA,
September 23 26, 1990.
28. VAN GOLF-RACHT, T.D., Fundamentals of Fractured Reservoir
Engineering; Elsevier Scientific Publishing Company, New York, NY,
pp. 179-180, 1982.

ProvenanceOriginal Petroleum Society manuscript, Sensitivity Study

of Coalbed Methane Production With Reservoir and Geomechanic
Coupling Simulation (2004-054), first presented at the 5th Canadian In-
ternational Petroleum Conference (the 55th Annual Technical Meeting of
the Petroleum Society), June 8 - 10, 2004, in Calgary, Alberta. Abstract
submitted for review November 28, 2003; editorial comments sent to the
author(s) May 18, 2005; revised manuscript received June 7, 2005; paper
approved for pre-press August 8, 2005; final approval September 22,
2005.

32 Journal of Canadian Petroleum Technology