SPE

Society of Petroleum Engineers

SPE 20729

Influence of Matrix Shrinkage and Compressibility on Gas

Production From Coalbed Methane Reservoirs
s. Harpalani, U. of Arizona, and R.A. Schraufnagel, * Gas Research Inst.
'SPE Member

Copyright 1990, Society of Petroleum Engineers Inc.

This paper was prepared for presentation at the 65th Annual Technical Conference and Exhibition of the Society of Petroleum Engineers held in New Orleans, LA, September 23--26, 1990.

This paper was selected for presentation by an SPE Program Committee following review.of information contained in an abstract sUb~ilted by the author(s). Contents of the paper,
as presented, have not been reviewed by the Society of Petroleum Engineers and are subJect to correction by the author(s). The matenal, as presen!ed: does not necessanly reflect
any position of the Society of Petroleum Engineers, its officers, or memberS. Papers presented at SPE meetings are subject to publication review by Edltonal Co~mlttees of the Society
of Petroleum Engineers. Permission to copy is restricted to an abstract of not more than 300 words. illustrations may not be copied. The abstract should contain conspicuous acknowledgment
of where and by whom the paper is presented. Write Publications Manager, SPE, P.O. Box 833836, Richardson, TX 75083-3836. Telex, 730989 SPEDAL.

ABSTRACT

The effect of methane desorption on gas flow through is in the form of compressed gas in pore spaces. In
coal was investigated. Adsorption/desorption isotherms coalbed reservoirs, the major quantity of gas is stored
were established for coal samples. Permeability of in adsorbed form, and the gas available for flow is the
cylindrical samples to different gases was measured at desorbing gas. The second maj or difference is the
simulated jn-situ conditions in a triaxial stress field. presence of the natural fracture (cleat) system in
Gas pressure-permeability-desorption relationship was coalbeds. Thus there are two main phenomena involved:
established for helium, methane and carbon dioXide. desorption followed by diffusion of gas towards the
Related experiments measured coal compressibility, and cleats, and ·Darcian flow of gas and water in the cleats.
volumetric changes of coal matrix with methane adsorptionFlow of methane in solid coal is subjected to two distinct
and desorption. processes: desorption of gas, and reduction in pressure.
Both these factors affect permeability of coal, and hence
The hypothesis that desorption of methane shrinks coal production of gas.
matrix, and results in increased permeability was
confirmed. The increase in matrix permeability with An experimental study - preliminary in nature - was
reduction in pressure from 1000 psi to atmospheric was carried out to investigate the effect of these two factors
highest for carbon dioxide (the most adsorptive gas), - desorption and pressure - flow characteristics of coal.
on the order of a factor of five for methane, and A brief description of the background that led to this
non-existent for helium. The corresponding volumetric study, experimental work, a discussion of the results
shrinkage of coal matrix on desorption of methane was and their significance, are presented in the following
0.6%. When measured values of compressibility and sections.
shrinkage were used as input parameters into a reservoir
simulation model, 60% more gas was produced over a five
year period. BACKGROUND

Decline in Pressure
INTRODUCTION
Once production from a coalbed methane reservoir starts,
Simulation of coalbed methane reservoirs and predicting the reservoir pressure starts declining. This decrease
long-term gas production is a complex process. The in pressure - although it takes place at a very slow
number of variables involved is large, and information rate - increases the effective stress. This increase
concerning these parameters and variation in their values in effective stress results in lower permeability. Stress
over the life of a producing reservoir is not well dependency of permeability has been studied by several
quantified. Knowledge of conventional gas reservoirs researchers in the past l • 2 • In general, compressibility
is not very useful due to the fundamental differences declines with increase in stress although there is
between sandstone reservoirs and coalbed methane res- insufficient data to explicitly define the relationshi p 3.
ervoirs. In conventional reservoirs, primary storage With large compressibility values such as those measured
at Rock Creek site in Alabama, a pressure drawdown of
only 300 psi - an increase in effective stress of 300
psi - can reduce the permeability by half or more 4 • This
References and Illustrations at end of paper dynamic changing of reservoir properties makes it
important to use proper pore volume compressibility

171
Page 2 ItlFLUENCEOF MATR IX SHR INKAGE AND COMPRESS IBILITY
ON GAS PRODUCTION FROM COALBED METHANE RESERVOIRS
values when attempting to analyze long-term production
data, or where reservoir drawdowns are large. It is
equally important to have a good understanding and
knowledge of how this value varies over the life of a
producing reservoir.
Another major impact of pressure variation on perme-
ability is the Klinkenberg effect increase in
permeability with decrease in pressure. The phenomenon
is ascribed to slip flow at the solid- gas boundaries
wi thin the pores and capillaries. When a gas is flowing
along a solid wall, the layer of gas next to the surface
is in motion with respect to the solid surface. In other
words, if the wall has a zero velocity, then the velocity
of the gas layer in the immediate vicinity of the wall
has a finite value, resulting in a higher gas flowrate
than expected5 • The application of Darcy's law to gas
flow in low permeability formations, therefore, requires
a correction for this gas slippage - or the Klinkenberg
- effectB. This correction takes the form of a slippage
factor, b, in the Klinkenberg equation.
............ ( 1 )
The apparent permeability, k, is approximately a linear
function of the reciprocal of mean pressure. The function
is an approximation, and the factor b is not a constant.
It increases with increasing gas pressure and decreases
with increasing water saturation5 ,B. Figure I shows the
Klinkenberg effect for a core sample of porous rock.
The intercept of the line on permeability axis indicates
the permeability of the media to a single incompressible
fluid.
Although the Klinkenberg effect is found to be true for
most rocks, several researchers have reported unusual
effects produced by coal. A non-linear relationship
between the two (reciprocal of mean pressure' and
permeability), and a negative slope have been reported 7 ,B.
Considering that a gas reservoir undergoes changes in
pressure and saturation during depletion, the effect of
slippage varies throughout the life of the reservoir.
It can have a significant positive impact on permeability
of coal and long-term gas production.
Desorption of Gas
The primary means of gas storage in coalbeds is by
adsorption of methane molecules on coal surfaces. When
subjected to pressure drawdown, desorbed gas moves by
diffusion through the extremely low permeability coal
matrix in order to reach the natural fracture (cleat)
system in coal. Once in the cleats, gas and formation
water flow in the cleat system which has a high
permeability relative to the matrix. Based on the above
mechanisms involved, several models have been developed
for predicting methane production from coal seams. An
excellent review of the various models has been presented
in recent publications 9 ,10,ll.
As production continues and reservoir pressure declines,
the cleats may contract - similar to contraction of pore
space in conventional reservoirs - causing a decrease
in permeability. On the other hand, the coal may contract
due to gas depletion causing an increase in permeabil-
ity 12. To date, enough information is not available to
determine the changes in permeability with decreasing
pressure. Production data, however, indicates that
permeability variations can not be fully explained by
172
:>t"'l: ZU/Z'j
the relative permeability effect alone. There appears
to be an additional parameter influencing permeability
of coal and its variation over the life of a producing
reservoir.
The purpose of this work was to investigate if there is,
in fact, a variation in the volume of coal in the matrix
as gas is depleted, and its impact on permeability. This
paper describes the experimental study carried out to
monitor variations in permeability with desorption, and
the corresponding volumetric strain of coal matrix.
Grain compressibility was also determined. Finally, the
results of a sensitivity analysis are presented where
the experimental results were used in a set of simulation
runs using a coalbed methane simulator to predict gas
production from coalbeds.
EXPERIMENTAL WORK
Specimen Preparation
Two primary types of experiments are discussed in this
paper: permeability experiments, and coal matrix volu-
metric s train experiments. For these experiments,
cylindrical specimens were needed. As a part of initial
sample characterization work, adsorption/desorption
experiments were conducted and proximate analysis of the
samples was carried out. For this part, powdered samples
were needed.
Fresh coal samples from gassy coal seams in the Piceance
and Black Warrior basins were obtained. Using these
lumps of coal, powdered coal samples were prepared in
the Mineral Processing Laboratory at the University of
Arizona. Pieces of coal were first crushed in a jaw
crusher to reduce the size to approximately one cm.
These were then ground in a ball mill for half an hour.
Powdered specimen was sieved into different size groups
and used for adsorption/desorption isotherms. To measure
permeability, and monitor the volume of the coal matrix,
specimens one and a half inches in diameter were used.
It was decided to keep the L/D ratio (L is the specime~
length, and D the diameter) as close to two as possible.
Hence the length of the specimen was close to three
inches. These specimens were prepared using a coring
machine and a surface grinder.
Adsorption Characteristics
In order to enable the flow of gas due to free gas alone
to be distinguished from the flow due to free and desorbing
gas, adsorption/desorption experiments were carried out.
Porosity of the sample was first determined using the
helium expansion method. Using the correction for gas
compressibility, the compressed gas component was
calculated. Following this, the adsorption isotherm was
established using methane. The procedure, used by the
authors in a previous study, is described in detail in
a prior publica~ion 13. After the highest pressure in
the adsorption experiment was reached, desorption
isotherm was established for decreasing gas pressure
from 1000 psi to atmospheric. Figure 2 shows the quantity
of gas adsorbed, along with the free gas content, for
one coal sample.
Since the primary objective of tllis research study was
to establish conclusive evidence of the relationship
between desorption and permeability, the effect of the
adsorptive affinity was investigated. Hence, carbon
dioxide was used as adsorbate and the experiment was
repeated. Carbon dioxide was selected because it is
SPE 20729 S. Harpalani and R.A. Schraufnagel
known to be more adsorptive than methane. However, the
experiment could be carried out only up to 500 psi.
Beyond this pressure, the results were difficult to
interpret. Either carbon dioxide liquefied, or there
was a chemical reaction between coal and carbon dioxide.
Figure 3 illustrates the difference between desorption
isotherms for carbon dioxide and methane.
Permeability Experiments
Permeability was measured using cylindrical specimens
at simulated in situ conditions in a triaxial stress
field. The experiments concentrated on studying the
flow behavior with changes in gas pressure (permeability
and pressure relationships). Since desorption takes
place when the gas pressure is reduced below the
desorption pressure, decreasing pressure - associated
with desorption - only is emphasized in this paper, alon~
with the Klinkenberg interpretation of the results (liP
vs permeability plots).
The experimental setup to measure permeability is shown
in Figure 4. A triaxial cell was modified by drilling
holes through the upper and lower platens to provide gas
inlet and exit ports. The synthetic rubber sleeve in
the cell sealed off the cylindrical surface of the coal,
from an annulus of oil which was pressurized using a
hydraulic pump. A compression testing machine was used
to apply the axial stress. Gas was supplied via a
regulator and pressure gauge to the top of the cell and
was collected at the downstream end. The flowrate was
measured using bubble flowmeters.
The specimen was inserted into the cell, and the setup
arranged as shown in the Figure 4. A small axial load,
and confining pressure was applied to the specimen to
seal the rubber membrane against the specimen surface.
Both stresses, axial and confining"were then increased
gradually and simultaneously until the desired stress
level was reached. The gas pressure at the upstream end
was then increased. Once flow of gas was established,
'the setup was left for several hours for the flow to
stabilize. After flow reached equilibrium, flowrate was
measured. Gas pressure at the inlet end was then increased
and the procedure repeated. After reaching approximately
1000 psi, the gas pressure was decreased in similar steps
until the flowrate became too low to be measured. The
stress was then released and the triaxial cell removed
from the testing machine.
To confirm that the observed changes in permeability
were, in fact, due to desorption of gas, pressure-
permeability relationships were established for three
different gases with significantly different adsorptive
characteristics. The experiment was first conducted
using helium since it is weakly adsorptive - almost
non-adsorptive. The specimen was then flushed with
methane for six hours and the experiment carried out
using methane. Finally, the experiment was repeated
using carbon dioxide - the highly adsorptive gas.
Matrix Volumetric Strain Experiments
The volumetric changes in the coal matrix, associated
wi th changes in gas pressure, were measured independently
of the permeability. It was necessary to keep the pressure
within the specimen, and all around it, equal at all
times to ensure that the void volume in the specimen
remained constant with changes in pore pressure. In
other words, pressure gradient had to be avoided for
these experiments. Following the standard procedures
173
Page 3
suggested by the International Society of Rock Mechanics
(ISRM) , four strain gauges were used on the specimen
surface - one each for axial and radial strains, 1800
apart. The strain gauge wires were passed between two
rubber a-rings, with the space in between completely
sealed with rubber cement. The specimen was then placed
in the sample container, the wires were connected to the
strain indicator, and the container was closed. The
setup is shown in Figure 5. The zero readings were
recorded and the specimen was evacuated for several
hours. Gas pressure was then increased to 100 psi with
helium. After the readings on the strain indicator
stabilized, they were recorded. The procedure was
repeated for pressure increments of 100 psi until 1000
psi was reached. Gas pressure was then decreased in
similar steps. At the end of this part of the experiment,
the specimen was evacuated and the entire procedure was
repeated using methane. After each pressure change the
apparatus was left for 8 to 10 hours, since desorption
is an extremely slow process and it took a long time for
the readings on the strain indicator to stabilize.
EXPERIMENTAL RESULTS
Permeability Variations
Using the measured flowrates during the experiments,
permeability of coal samples was calculated using Darcy's
law for steady state isothermal flow of a gas through
porous media. A complete gas pressure-permeability
relationship was thus established. Figure 6 shows the
results of one experiment carried out at a hydrostatic
stress of 1500 psi (10.34 MPa). Permeability started
increasing below 450 psi (3.1 MPa) and continued to rise
sharply until the pressure dropped to 50 psi (0.34 MPa)
'and the experiment was complete.
As mentioned in the previous section, the effect of
desorption of methane on permeability of coal was further
confirmed by using different gases, with substantially
different adsorptive affinity for coal. Figure 7 shows
the variations in permeability of the sample to the three
gases used for decreasing gas pressures. These results
clearly point out that the permeability increase with
reducing pressure is highest for carbon dioxide, on the
order of a factor of five for methane, and non-existent
for helium, suggesting that the change in permeability
is not only a function of desorption, but the degree of
desorption as well.
Coal Matrix Volume Variations
Figure 8 shows the changes in volumetric strain (t>V IV)
with increasing and decreasing pressure for helium and
methane. For helium, the volume of coal matrix decreased
with increasing gas pressure. Grain compressibility of
coal, as calculated from Figure 8, is approximately 9
x 10- 10 per psi. The hysteresis between increasing and
decreasing gas pressures is due to the slightly plastic
nature of coal. With methane, the matrix volume increased
linearly with pressure. At 1000 psi, the matrix volume
was swollen by 0.6%. When gas pressure was reduced, the
decrease in volume was non-linear, and there was a
residual swelling when the pressure was reduced to
atmospheric.
Page 4 INFLUENCE Of MATKIX
ON GAS PRODI,JCTI ON FROM COALBED ~IETHANE RESERVO IRS
DISCUSSION OF RESULTS
Effect of Desorption
Figure 9 demonstrates the relationship between desorption
and permeability for methane. Once desorption - which
is significant below 600 psi (4.14 MPa) starts,
permeability starts increasing. The initial decrease
in permeability above the desorption pressure - prior
to desorption - is due to cleat compression which impedes
free gas flow. Figure 10 shows the same relationship
for carbon dioxide. Since carbon dioxide is highly
adsorptive, the permeability rises continuously as the
gas pressure decreases. The initial decrease in
permeability to carbon dioxide is not observed because
the maximum gas pressure (of a little less than 600 psi)
is significantly below the desorption pressure, and the
flow of carbon dioxide is accompanied by desorption for
the entire experiment. The results also indicate that
the increase in permeability of coal to carbon dioxide
is about 30 times, whereas for methane it is approximately
6 times - even though the measured permeability vlues
are fairly low.
To further confirm that the observed increase in
permeability was due to desorption, one set of experiments
was carried out using a back pressure at the downstream
end of the specimen to minimize pressure gradients across
the sample. The stepwise gas pressure experiment was
then repeated at a constant gas pressure gradient by
changing the pressure at both ends. ~ was constant
throughout the experiment, and mean gas pressure changed,
thus giving a mean gas pressure-permeability relation-
ship. In sjtu conditions were better simulated since
very high pressure gradients - as used in the prior
experiments - are not found in reservoirs even if large
gradients occur near drawndown production wells. Figure
11 shows the results for pressure gradients of 300, 200,
100 and 50 psi. As the gradient was reduced, an
everincreasing permeability with reduced mean gas
pressure below the desorption pressure was observed.
This confirms the effect of desorption on permeability.
In order to obtain the effect of desorption on matrix
volume of coal, the volumetric strain wi th helium (grain
compressibility) was subtracted from the volumetric
strain with methane. Figure 12 shows this as 'effective
shrinkage' due to desorption. The similarity between
the desorption isotherm and the effective shrinkage plot
suggests a clear interdependence.
Numerical Analysis
Curve fitting exercise was carried out using the
experimental results. The variation in permeability
with decreasing gas pressure, shown in Figure 6, was
taken to be the standard relationship. The best fit
gave an equation of the form
k=3.3+'076+'0206 p2 ............ (2)
P
where P is the gas pressure in 100 psi. Figure 13
compares the modeled results with the experimental
results. The two are in good agreement. Similar curve
fitting for other results gave non-linear, and similar,
equations with three coefficients.
~HKINKA~t ANU ~UM~Kt~~IDILI IT
The interesting feature of the above equation is its
similarity to the Klinkenberg Equation, with one extra
term. At low gas pressure, the first term in the equation
dominates although this is the combined effect of
desorption and Klinkenberg phenomena. At high gas
pressures, the third term is dominant resulting in higher
permeability. To better understand the relationship,
the results were plotted as a function of liP, where P
is the mean gas pressure. Figure 14 shows the plots for
two experiments carried out at hydrostatic stress levels
of 1100 psi and 2000 psi. For the 1100 psi experiment,
there is a minimum permeability, with its value increasing
in either direction - increasing and decreasing pressure.
This might be due to variation in the value of the
slippage factor b. At 2000 psi, the sample behaves more
like a typical rock with the permeability rising
continuously as gas pressure is reduced - except for the
initial flat portion of the plot. The initial deviation
from the usual Klinkenberg phenomenon is minimized and
the flow appears to be primarily capillary flow. Although
experiments were not carried out at stresses higher than
2000 psi, it is felt that at high stresses the permeability
would increase continuously with reductions in gas
pressure, i. e., the k v s liP line would be a straight
one, at say, 3000 psi. Since pressure in coalbed methane
reservoirs decreases with depletion of gas, the perme-
ability might go up continuously for reservoirs with
high jn-situ stresses. If this can be shown/proved to
be true, the phenomenon would favor methane production
from deep coalbeds, e.g., Piceance basin.
Importance of Coal Matrix Shrinkage
It has been clearly demonstrated in the preceding sections
that there is a matrix shrinkage phenomenon associated
with desorption of methane. An outcome of this has been
observed in the experimental results where permeability
increases when accompanied by desorption. In this
section, an analysis is presented to evaluate the
sensitivity of gas production to coal matrix shrinkage
using the results of the experimental work in a set of
simulation runs using COMETPC-3D model. Four simulation
runs were made to determine the effect of pore volume
and matrix compressibilities on five year gas production.
Measured value of matrix (or grain) compressibility was
9xlO- IO per psi. Hence, C p =9xlO- 7 psi-'. Volumetric
measurements using methane were incorporated in the
simulation runs using the concept of a matrix shrinkage
compressibili ty, C~14.
......... (3)
From the results, the volume of matrix decreased by 0.62%
when gas pressure was reduced by 1000 psi. Therefore,
c' =0.0062=6.2xlO- 6 ·-1 ........ ( 4)
m 1000 pSI
Unfortunately, pore volume and bulk compressibilities
were not measured as a part of this experimental study.
Hence, the pore volume compressibility had to be
estimated. Using the common assumptions in reservoir
engineering, even though these have not been proved for
coalbed methane reservoirs, its value was approximated
to be 4.Sx 1 0'5 psi -I, which is a reasonable value 15 • These
values of compressibilities, as used in the simulation
runs, are shown in Table 1. The results of the simulation
runs are shown in Figure 15.
174
SPE 20729 S. Harpalani and R.A. Schraufnagel
Case 1, where C p = C;" = 0, is used as the basis of
comparison. A comparison with the results of Case 2
suggests that production can be underestimated by
considering the pore volume compressibility and not the
matrix shrinkage compressibility. Case 3 shows that
matrix shrinkage compressibility can more than offset
the reduction in permeability due to pore volume
compressibility, and the resulting increase in production
can be fairly significant. Coal matrix shrinkage is,
therefore, an important parameter and must be considered
together with cleat contraction due to pore volume
compressibili ty in order to accurately simulate and
forecast long-term gas production from coalbed methane
reservoirs. However, it should be pointed out that more
laboratory work needs to be carried out in order to
quantify both pore volume and matrix shrinkage com-
pressibilities.
CONCLUSIONS
The most important conclusion of this work is that
desorption of gas causes the coal to shrink. The shrinkage
trend with decreasing pressure parallels that of the
desorption trend. The volume of coal matrix shrinks by
approximately 0.5% with a reduction in pressure from 600
psi to atmospheric. If the compressibility of grains
is taken into consideration, the effective matrix
shrinkage compressibility is even higher. For simulation
purposes, C;" was taken to be 6.2xlO' o psi-] since volume
increase was ,62% at 1000 psi. If grain compressibility
C m (-9xlO- 7 ) is included, the
C~=(6.2xlO-o+9xlO-7) psi-' or 7.1xlO- o psi-] and the
effect shown would be even more significant.
The shrinkage of coal matrix causes the permeability to
increase significantly once the pressure falls below the
desorption pressure, thus resulting in higher production
of gas. The effect of shrinkage on permeability can be
easily visualized by the gas pressure-permeability
relationship for pressure reduction from 1000 to 50 psi.
Above the desorption pressure, the effect of pore volume
compressibility is dominant and the permeability
decreases. However, once the pressure falls below the
desorption pressure, the effect of matrix shrinkage
becomes dominant resulting in increased permeability.
The relative change in permeability with desorption is
strongly dependent on the adsorptive affinity of gas for
coal. It is higher for carbon dioxide than for methane,
and non-existent for helium.
The influence of desorption and Klinkenberg effect on
permeability of coal is complex. In this study, the two
effects could not be separated. At high stresses, the
results appear to favor the Klinkenberg effect.
Also, interesting is the observation that permeability
is much higher at low pressure gradients. Lowering
pressure gradients even further might further increase
permeability and show that desorption plays a key role
in maintaining, and potentially increasing, permeability
during the production life of a well.
Future Work
Based on the findings of this' exploratory' work, further
research is desirable in the following areas: .6
Dual-Mechanism Approach to the Flow of Gas in Tight
1. Analyze additional samples from basins where coalbed Formations," SPE Formation Evaluation (1986) 43-52.
methane is an important resource. These experiments
175
Page 5
should be conducted using large samples, at least three
inches in diameter - preferably four inches. Also,
transient testing should be conducted (rather than the
steady state tests) for permeability measurements.
2. Experiments should be conducted to establish a range
of compressibility coefficients - pore volume, bulk,
matrix and matrix shrinkage - for different coal basins
in the U.S.
3. Klinkenberg phenomenon should be studied in detail
in order to isolate·the change in permeability due to
the Klinkenberg effect, and that due to desorption/
shrinkage. This is particularly important for methane
recovery from deep coalbeds - systems already developed,
as well as for future operations.
NOMENCLATURE
k permeability
b slippage factor
P mean pressure
Cp - pore volume compressibility
Cm - matrix (grain) compressibility
Cm* ~ matrix shrinkage compressibility
Vm ~ specimen volume
~Vm - change in the specimen volume
6P - pressure change
00 - infinite pressure
ACKNOWLEDGMENTS
This work was carried out as a part of the GRI sponsored
project "Permeability Changes Resulting from Gas
Desorption," Contract No. 5088-215-1666. The authors
wish to thank the GRI for the support. The simulation
work using COMETPC-3D was carried out by Walter K. Sawyer,
and the authors are thankful to him for his assistance.
REFERENCES
1 Harpalani, S. and McPherson, M.J.: "Effect of
Stress on Permeability of Coal," Quarterly Review of
Methane from Coal Seams Technology, Vol. 3, No.2 (1985),
23-28.
2 Somerton, W.H. et al: "Effect of Stress on
Permeability of Coal," Int. J. of Rock Mechanics, Min.
Sci. and Geo. Abstr., Vol. 12 (1975) 129-145.
3 Koenig, R.A. et al: "Application of Hydrology to
Evaluation of Coalbed Methane Reservoirs," final report
submitted to Gas Research Institute, Contract No.
5087-214-1489 (1989).
4 Zuber, M.D. et al: "The Use of Simulation and
History Matching to Determine Critical Coalbed Methane
Reservoir Properties," SPE/DOE 16420 paper presented
at the 1987 Low Permeability Reservoirs Symposium,
Denver.
5 Klinkenberg, L.J.: "The Permeability of Porous
Media to Liquids and Gases," Drilling and Production
Practice, API (1941) 200-213.
Ertekin, T. et al: "Dynamic Gas Slippage: A Unique
Page 6 INFLUENCE OF MATRIX SHRINKAGE AND COMPRESSIBILITY ~ n. "LU I"L 'j

ON GAS PRODUCTION FROM COALBED METHANE RESERVOIRS

7 Gawuga, J.: "Flow of gas through stressed .11 Sawyer, W.K. et al: "Development: and Application
carboniferous strata," Ph.D. Dissertation, University 'of a 3D Coalbed Simulator," CIM/SPE 90-119 paper presented
of Nottingham (1979). at CIM/SPE IntI. Tech. Meeting, Calgary, JcU1e 10-13.

8 Harpalani, S.: "Effect of Gas Pressure on
Permeability of Coal," Proceedings, 2nd U. S. Mine
Ventilation Symposium, Reno, Sept. 23-25, 1985.
9 King, G.R. and Erteken, T.: "A Survey of
Mathematical Models Related to Methane Production from
Coal Seams, Part 1: Empirical and Equilibrium Sorption
Models," Proceedings, 1989 Coalbed Methane Symposium,
Tuscaloosa, April 17-19, 125-138.
10 King, G.R. and Erteken, T.:
Mathematical Models Related to Methane Production from
Coal Seams, Part II: Non-Equilibrium Sorption Models,"
Proceedings, 1989 Coalbed Methane Symposium, Tuscaloosa,
April 17-19, 139-156.
12 Kamal, M.M. and Six, J .L.: "Pressure Transient
Testing of Methane Producing Coalbeds," SPE 19789 paper
presented at the 64th Annual Technical Conference and
Exhibition, San Antonio, October 8-11.
13 Harpalani, S. and McPherson, M.J.: "Retentionand
Release of Methane in Underground Coal Workings," Int.
J. of Min. and Geol. Engr., Vol. 4 (1986), 217-233.
14 Sawyer, W.K.: Personal Communication, 1989.
IIA Survey of
15 Harpalani, S.: "Permeability Changes Resulting
from Gas Desorption," final report submitted to Gas
Research Institute, Contract No. 5088-215-1666 (1989).

TABLE 1 - PARAMETERS USED IN SIMULATION RUNS AND RESULTS

Case Cp(psi- 1 ) C~(pSi-l) Gas Prodn. Change

1 0.0 0.0 80,400 0.0

(base case)

2 4.5 x 10- 5 0.0 69,100 -14.1%

3 4.5 x 10- 5 6.2 X 10- 6 105,800 31.6%

4 0 6.2 X 10- 6 127,900 59.1%

fOe 1/11/
f5 e /Vl /

~4.0 V
i I /i?~~~~~--;:::±=:=====±:,-1
35 e
1

.•, /
I I

~ 3J~
::
~~ : :::::::;n l!J. Corbon dioxide
e-

Vl

<3 _L-_,L..-....L-"L-.,...L----l_::'"'=---l'_-::'I:::--::'I~:_l::_---l~:_"I:_'
2.5 L..-.....L.'
0.2 0.4 0.6 0.8 1.0 1.2 1.4
Reciprocal mean pressure, ATM- 1

Figure 1. Permeability of core sample to three gases

at different gas pressures.

176
 SP.E 2072,9

28
28
Total Gas

24 24
E
E '"
"-
'"
"-
20 E 20
E Adsorbed Gas
-a.
.
.; ,Q
16
Cl 16 0
.
~
. '". .
• Methane
E "
E
~
,: .It.. Carbon dioxide
0 0
> >

Compressed Gas

a 100 200 300 400 500 600 700 800 900 1000
a 100 200 3 a 400 500 600 700 800 900 1000
Gas Pressure, psi
Gas Pressure, psi
Figure 3. Desorption isotherms for the coal sample using methane and
carbon dioxide, 2SoC.
Figure 2. Capacity of coal to hold methane at 25 (Ie (dry sample).

Pressure gauge

/ Valve

To strain

indicator

From strain

gauges #2
Go>
Specimen

Hydraulict
System ~ Axial strain Radial strain

gauge #1 gauge #1
Figure 4. Schematic diagram of apparatus for
testing coal permeability.
Figure 5. Setup to monitor changes in the volume of
coal matrix with variations in gas pressure.

7
Hydrostatic Stess: 1500 psi

'U increasing gas pressure

.: 5
'"
N decreasing gas pressure

..,E
0 4
';;

~
.
:<i 3
"E
;;
ll. 2

a 200 400 600 800 1000

Gas Pressure, psi

Figure 6. Variation in permeability of coal to

methane with increasing and decreasing gas pressure.

177
 SPE 2072,9

hydrostatic stress: 1500 psi .006

Experiment # 2
helium
.005
~ Mathane cycle
methane .004

:;
.
<l

.OOJ
Experimenl # 1

E- "u Adsorption
N

-.
E
;:;
",
0 D

g
.
f~:::::::~:i;4~0~0~;~6~0~0:::~800:==1:0~0°C:i
:;;
.1
"
E
Q .08
carbon dioxide
-.001 Helium cycle
... .06 Gas PrBSSUrB. psi
.04
Figure 8. Volumetric strain of coal matrix for increasing and
decreasing gas pressure for helium and methane .
.02

.01 0 200 400 600 800 1000

Gas Pressure, psi

Figure 7. Variation in permeability of coal

to helium, methane and carbon dioxide for
decreasing gas pressure.

7 .... adsorbed gas (28-48 mesh)

• adsorbed gas (28-48 mesh)
• permeability (hydrostatic
• permeability (hydrostatic
6 J.O stress: 1500 psi)
stress: 1500 psi)
:;
.., 5 25 2- 2.5 25
E- N

N
E
E E E
", '"
,.
"
0
4 '"
"-
E
0 2.0
"-
E
20

';; .; .;
" ~ "
~
.tl

...
J .tl 15
..,...
:;; 1.5
..,.
0

.
:;; 0

. E
"
E
Q 2 ...Q "
" 1.0 E 10
... E
.=0 .!
0
> >
.5

o o 400 600 8 1000 0

Gas Pressure I psi Gae Pressure, psi

FiRure 9. Variation in adsorbed gas and permeability Figure 10. Variation in adsorbed gas and permeability of
of coal to methane - decreasing pressure only. coal to carbon dioxide - decreasing pressure only.

178
 SPE 20 72.9

Hydroslalic siress: 1500 psi .006

Efhclil8 shrinkage
20
.005 Mllhane
-g, 10 ~

8
<i
,. 6
.
o 4 ~
u

'"
D
.'"
2
.
j
>
.001
E
.. I 0
a..
...;2=<OO;.::..-e.......:.4..;OO-=--.-_6~0~0~_~8:0~O_ _=1000
.8
.6
• ap=300 psi
-.001
r Helium

Gas pressure. psi

Mean Gas Pressure, psi Figure 12. Volumetric strain of coal matrix for decreasing gas pressure.

Figure 11. Variation in permeability of coal to

methane with mean gas pressure, at constant
pressure gradients.

8-,----------------,
10
7

~ 8 1130 psi
~6 'K - 3.2692/P + 0.0763 + 0.0206P"
o N
'tJ
~s .,E
u
'E 0

-4
g .;.
=
~
:=3
J:l
D

.'" 2000 psi

o
III

E2
III
.~
a..
a..

0+--,.-.--..--,--,--.--,--,---,.-.---1 .04
o 1 2 3 4 5 6 7 8 9 10 11
1/PfIII pSi-I
Gas Pressure Gradient(P), x100 psi
Figure 13. Experimental and modeled results Figure 14. Variations in permeability of coal
for variations in permeability of coal to with reciprocal of mean gas pressure.
methane.

100

90

POSITIVE EFFECT OF
MATRIX SHRINKAGE
":;:
4-

::E: 70
.;
+'

'"'"
<::
0 NEGATIVE EFFECT OF
:;:; 50 PORE VOLUME
u
COflPRESSIBILITY
"'"s-
0
o..
~
30
'"
<.0

10
o 4
TIME, YEARS
Figure 15. Effect of compressibilities on gas production.

179