International Journal of Mining and Geological Engineering, 1990, 8, 369-384
Measurement of parameters impacting methane recovery from coal seams S. H A R P A L A N I * ~ and A. S C H R A U F N A G E L § ~;Department of Minin 9 and Geological Engineering, The University of Arizona, Tucson, Arizona 85721, USA §Gas Research Institute, 8600 West Bryn Mawr Avenue, Chicago, Illinois 60631, USA
Received 4 January 1990
Summary This paper describes the behaviour of coalbeds as gas reservoirs and discusses the results of a study carried out to establish the effect of release of methane on gas flow behaviour of coal. Experimental work consisted of microscopy, establishing adsorption/desorption isotherms, and monitoring changes in the volume of coal matrix with increasing and decreasing gas pressure. Micrographs obtained using small pieces of coal indicated that coal is made up of blocks, containing matrix and pores, separated by microfractures. This confirms the dual porosity model of coal structure with a primary porosity, and a fracture/cleat porosity-physical model used in coalbed methane simulators developed recently. Isotherms suggested that for the samples tested, a major part of the gas is released only after pressure falls below 600 psi, and this is primarily due to desorbing gas. Results of the volumetric strain experiments indicated that there is an increase in matrix volume with increase in gas pressure, in spite of matrix compressibility. Adsorption, therefore, induces swelling of the matrix. With decrease in gas pressure from 1000 psi to atmospheric, the matrix volume shrunk by 0.5%. These experimental results were inputted in a reservoir model and simulation runs made to determine the effect of pore volume and matrix shrinkage compressibilities on gas production. Over a five year period 60% more gas was produced when matrix shrinkage was used as an input parameter.
Introduction Recently the US Geological Survey reduced by 40% its estimate of undiscovered domestic sources of conventional oil and gas. Although these projections are subject to refinement, there can be little doubt that unconventional resources will play a vital role in meeting the US energy demand in the years to come. One such resource is coalbed methane, contained in large quantities in coal seams. Liberations of methane in underground coal mines have always been a major concern to coal operators because of the risk of explosion associated with ignition of methane. Usually, Editor's note: The units used in this paper are generally used by the Gas Research Institute and are found in most oil and gas publications. Conversions of the more important units are: 1 MMCFD---28 300 m3/day; 1 MSCFD_= 28.3 m3/day; 100 psi~-0.68 MPa; 100 ft2/lb~46.87 mZ/kg; 1 ft---0.3048 m; 1 acre---0.40 hectare. *Author to whom all correspondence should be directed. 0269 0136/90 $03.00+.12
© 1990 Chapman and Hall Ltd.
370
Harpalani and Schraufnagel
this has been dealt with by increasing airflows to reduce concentrations of methane in mine air and by venting the gas to the atmosphere. Some deep mines produce 10 to 15 million ft3/day ( M M C F D ) of methane and require circulation of as much as 10 to 15 tons of air per ton of coal mined in order to clear the gas from the mine (Cervik, 1967). In recent years, effort has been made to convert this liability to an asset by recovering the gas prior to mining, as well as from unminable coal seams. Although the interest started off in order to ensure safety in underground coal mines, coalbed methane is now becoming a valuable source of energy. Gas production from coal seams in the US increased from less than 1 M M C F D in 1980 to more than 130 M M C F D in 1988 and had grown to over 300 M M C F D at the end of 1989. Active development is underway in Australia and interest is growing in Europe. Gas from coalbeds is beginning to represent a significant resource base that is expected to grow in the future (Kamal and Six, 1989). During the last few years significant advances have been made in defining coalbeds suitable for methane recovery, estimating the gas content of potential reservoirs, and designing systems for efficient and effective recovery of this resource. Studies have been carried out to determine flow characteristics of coal and variation in flow behaviour over the life of producing wells. Several factors have been found to affect permeability of coal and hence production from coalbed methane reservoirs. Presence of water and initial dewatering of coalbeds, in situ stress - particularly in the case of deep reservoirs - reservoir pressure, stimulation techniques, are among the primary parameters being studied in detail. Recently, another factor has been found significantly to affect coal permeability: shrinkage of solid coal as a result of release of gas. Preliminary study indicates that this can have a major impact on gas production from coalbed methane reservoirs. This paper discusses the general gas flow behaviour in coalbeds, and describes the experimental study carried out to estimate the matrix shrinkage due to degasification. Finally, the influence of shrinkage on gas production is discussed, along with the results of the simulation runs carried out using the experimental results as input.
Background Reservoir description Coal differs from other types of sedimentary gas reservoir rock materials in that the gas is stored as an adsorbed phase on walls of the coal micropores. Since the internal surface area of coal can be as large as 100 000 ftZ/lb of coal, the quantity of adsorbed gas is extremely large (Jones and Bell, 1987). It is common to find coalbeds which can store 3 to 7 times the amount of methane per cubic foot of reservoir rock that conventional sandstone reservoir can store at similar depths and pressures. The sandstone reservoir releases its gas uniformly, since it consists of matrix and pore space. The gas is stored in the pore space and the gas-in-place is a function of formation porosity, pressure, temperature and water saturation (Kamal and Six, 1989). There is usually little or no mobile water associated with conventional gas reservoirs, and relative permeability is not a significant factor in managing these reservoirs. When production begins, and the reservoir pressure declines, little or no changes are observed in the formation permeability. In contrast, gas in coal reservoirs is released in a highly nonlinear manner and the major fraction of adsorbed gas is released at low pressures. The difference between the two types of reservoirs is shown in Fig. 1. In order to recover a large
Measurement of parameters impacting methane recovery from coal seams
371
X~ L.
0 ul
/ /
/
/
/ Sandstone
fi
i I
/// PIZ
Fig. 1. Comparison of gas storage for coal with an equivalent capacity sandstone (Jones and Bell, 1987) percentage of gas from a coalbed, the reservoir pressure must be reduced significantly. This is usually achieved by removal of water for long periods of time. Another difference between conventional gas reservoirs and coalbed methane reservoirs is that coal seams are generally characterized by dual porosity, containing both a micropore porosity (primary porosity) and a macropore porosity (fracture porosity) system. The micropores have a dimension ranging from 5 to 10 A and exist in the coal matrix between the seam's cleats. The macropore system is made up of the volume occupied by the cleats of widths varying from a few ~ to microns (King, 1985). The model commonly used is shown in Fig. 2, where S is the fracture spacing. Most numerical and computer models developed recently to simulate coalbed methane reservoirs are based on this type of physical structure of coal. A coalbed is shown to be made up of small blocks separated by cleats. The face cleats are continuous, while the butt cleats are normally discontinuous intersecting the face cleats at right angles. Due to the dimension of the pores the micropore system is inaccessible to water. Hence this serves as the storehouse for large quantities of methane held within the porous matrix by an adsorption mechanism which is controlled by the reservoir pressure. The macropore system, on the other hand, is water saturated prior to production. When a coalbed methane well begins to operate, it is common to encounter only single-phase flow, i.e. only water is produced (Koenig et al., 1989). As water production continues and the hydrostatic pressure is reduced, coal becomes less capable of retaining methane which then begins to desorb from the surface of the coal matrix. Desorbed gas causes the concentration of methane within the matrix to go up resulting in a concentration gradient and consequently in diffusion of gas through the matrix towards the cleats. Gas bubbles form in the pores and the cleats, coal is now partially saturated with gas and the relative permeability of coal to water decreases. With continued pumping, a two-phase flow regime is reached. The gas saturation increases and the gas begins to flow too. This sequence of three flow regimes, shown in Fig. 3, moves outward from the wellbore into the formation over time. Relative permeability is therefore one of the most important parameters affecting production from coalbed methane reservoirs. Extensive experimental studies have been carried out to investigate the relative permeability
Harpalani and Schraufnagel
372 Butt
,.
cleat
¢
;,
~ ~
'~-.
•~
•
,# b ~ ~.y .," ,~*
u ...r._~,~.:' .- :..~, ,~. . . . ' ~ ^ ~ i_Z_' ~ " ~ . . , , ~ ' - , ~
i•~,-,' + 0 "~'t~ "~+
.- " "~J ":<~'~; ~;+~ i~i
• "- "0",~',~::<..~, ' ,,o,t ..<~. -. O:
,:..+
+
+ .
"O
+
.o,~L
Matrix
blocks
:~
;,o,.~'., d.
". ~.+.~_~'~,o,~.-,-
.~+,vj-"~ . . ~ + 1"~.; .+'~+.
t
,- •+.:,"
"~" :~."I+'~ ~ ".+ -e'~ ,o-., +j, .~+.,. L~
~3+. containing
,,,,+~...
;;.;'+ ~
# .~ ..
+.:
micropores
Fig. 2. I l l u s t r a t i o n of coal fracture system ( S a w y e r et al., 1987)
Distance from well
¢
~>~ 1.O
Stage 3 Two- p h a s e f l o w regime
Stage 2 Unsaturated f l o w regime
I
Stage 1 Saturated f l o w regime
I
E o.
a=
o.o
Fig. 3. T h r e e s t a g e s o f c o a l b e d m e t h a n e p r o d u c t i o n
(Koenig
et al.,
1989)
Measurement of parameters impacting methane recovery from coal seams
373
of coal (Reznik et al,, 1974; Taber, 1974). A wide variation (<0.01 to 100 md) in air and water permeability was measured for each type of coal in the study. This is often the case due to the nonhomogeneous nature of coal. However, the flow was found to be primarily through microfractures (Dabbous et al., 1974), and this is in agreement with the dual-porosity models used in simulation of coalbed methane reservoirs. An interesting feature was that although the measured permeability values varied significantly, the general behaviour for all types of coal was always very similar. Dewatering and degasification of a coalbed results in an overall decrease in fluid pressure in the reservoir, resulting in an increase in effective stress. Increased stress is known to decrease permeability due to compaction. This phenomenon has been studied by several researchers in the past (Reznik, 1974; Somerton, 1975; Harpalani, 1985). In general, increase in stress decreases the compressibility although there is insufficient data to define explicitly the relationship (Koenig, 1989). With large compressibility values, such as those measured at the 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 (Zuber et al., 1987). From the above discussion, it is apparent that the important parameters determining flow behaviour in coalbed methane reservoirs are relative permeability, stress, gas pressure - all of these being dynamic, changing continuously with time - adsorption characteristics, and the microstructure of coal. For some time now another factor is suspected of influencing the permeability of coal: shrinkage of solid coal as a result of desorption of methane. It is common knowledge that removal of bound or collodial water can dramatically change properties of the solid. Studies reported in Gregg (1961) support the theory that desorption of gas shrinks the solid adsorbent due to an increase in free surface energy of the adsorbent. Moreover, there is evidence from production that coal permeability can increase over the life of a gas reservoir increasing its ability to produce gas - an increase that cannot be explained by conventional reservoir engineering theories.
Purpose of this study Considering the level of knowledge relating desorption, associated shrinkage of coal matrix and the resulting increase in permeability, it was felt that an experimental study should be carried out to investigate the relationship between the three, and the impact of this on gas production. The main purpose of this experimental investigation was to p r o v e - o r disprove - that permeability can increase when gas is desorbed. Prior to experimental work to measure the volumetric changes in coal matrix and permeability, adsorption/desorption characteristics of coal were established. Also, the microstructure of coal was investigated using a scanning electron microscope (SEM). The following sections discuss this initial characterization work and the experimental work to measure permeability and matrix shrinkage. Finally, an analysis to estimate the impact of this additional parameter on production was carried out using the experimental results as input, and running simulation runs using a coalbed methane simulator.
Experimental work Adsorption/desorption isotherms Using the indirect method of gas expansion, an adsorption isotherm was first established for pressure up to 1000 psi. The procedure is described in detail in a prior publication
Harpalani and Schraufnagel
374
(Harpalani and McPherson, 1986). Following this, the desorption isotherm was established for decreasing gas pressure. Since the desorption isotherm is of primary interest in coalbed methane reservoir engineering, this alone is shown in Fig. 4. The isotherm indicates that a major part of the gas is released only after the pressure falls below 600 psi. Hence, gas flow below this pressure will be accompanied by desorption.
28
24
20
12
0
I 0
I
i
i
i
I
i
I
I
200
300
400
600
600
700
800
900
Gas pressure ( p s i )
Fig. 4. Adsorption isotherm for the coal sample (dry sample, moisture-freegas)
Microstructure of coal Small chunks of coal were taken and glued to the specimen holder using colloidal carbon. The specimen was then coated with gold-palladium to minimize charging of the sample by the electron beam. The specimen was now ready for analysis using an SEM (ISI-Super IIIA). Since no specimen preparation - grinding, polishing, etc - is necessary for SEM, the samples were completely undamaged making it possible to obtain specific information concerning the pores and fractures on the specimen surface - as seen in the micrographs. Figure 5 is a micrograph showing size and distribution of pores on the surface of a piece of coal. A large number of randomly distributed pores can be seen. Figure 6 is another micrograph showing the surface configuration of the sample. It shows the coal structure as made up of 'fibres' going in different directions. Pores probably exist but were difficult to identify at points where fibres meet or are broken. The surface area is indicated once again to be very large due to the uneven surface. Figure 7 shows the networking of microfraetures on the surface of the coal sample indicating the validity of the dual-porosity model discussed earlier.
Measurement of parameters impacting methane recoveryfrom coal seams
375
Fig. 5. SEM micrograph showing pores on the surface of the specimen ( x 2300)
Shrinkage/swellin9 experiments The volumetric changes in the coal matrix associated with adsorption and desorption were measured independently of the permeability measurements. Determination of permeability involves maintaining a pressure gradient across the specimen, resulting in a change in the volume of the pores, fractures and microcracks within. The measured volumetric changes would therefore be due to the sum of changes in the matrix volume, as well as void volume. It was therefore necessary to keep the pressure around the specimen equal on all sides. In the absence of a pressure difference between the pore pressure and outside pressure, the volume of the voids in the specimen would remain unchanged. Measured volumetric changes would thus be due to changes in the volume of the coal matrix alone. Using cylindrical specimens, 1.5 in. in diameter and 3 in. long, and following the standard procedure suggested by the International Society of Rock Mechanics, four strain gauges were used on the specimen surface - one each for axial and radial strains, 180 ° apart. The strain gauge wires were passed between two rubber O-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 a strain indicator and the container was closed. The set-up is shown in Fig. 8. The zero readings were recorded and the specimen was evacuated for several hours. Using helium, the gas pressure was increased to 100 psi (0.69 MPa). After the readings on the strain indicator stabilized, they were recorded. The procedure was repeated for pressure increments of 100 psi (0.69 MPa) until the pressure reached 1000 psi (6.9 MPa). 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
Harpalani and Schraufnagel
376
Fig. 6. SEM micrograph showing the surface of the specimen ( x 2000)
methane. After each pressure change, the set-up was left for 8 to 10 h to equilibrate since adsorption/desorption is a slow process. The average of the two measured axial strains was taken as the net strain in the axial direction. This was repeated for the radial direction as well. Volumetric changes were thus calculated for each pressure level, and the volumetric strain AV/V was determined.
Results and discussion
Volumetric strain Figure 9 shows the changes in volumetric strain (A V/V) with increasing and decreasing gas pressure for helium and methane. For helium, the volume of coal matrix decreased with increasing gas pressure. At 1000 psi (6.9 MPa), the volume reduced by 0.087%. Grain compressibility of coal is, therefore, 8.7 x 10 -7 per psi (1.26 x 10 -1° per Pa). When the pressure was decreased, the volume did not return to its original value. There was a net decrease in specimen volume of 0.026% because coal is not perfectly elastic. When the experiment was repeated using methane, the volume increased linearly with each increase in gas pressure. At 1000 psi, the volume of coal matrix increased by approximately 0.5%. Therefore, adsorption induces swelling of the coal matrix. When the gas pressure was reduced, the decrease in the matrix volume was nonlinear. At zero pressure, the volume of the specimen remained 0.1% higher than its orginal value. This indicates either that more time
Measurement of parameters impacting methane recovery from coal seams
377
Fig. 7. SEM micrograph showing microfractures on the specimen surface ( x 900) was required for complete desorption or that coal permanently acquired a residual gas content. Figure 10 shows the volumetric strains for decreasing pressure, i.e. for desorption only. To obtain the effect of desorption alone, the volumetric strain with helium (grain compressibility) was subtracted from the volumetric strain with methane. This respresents the effect of desorption on the volume of the coal matrix, and is shown as the uppermost curve in Fig. 10. It is apparent that the shape of the 'effective shrinkage' plot is very similar to that of the desorption isotherm. The two are plotted together in Fig. 11, clearly indicating that shrinkage is directly dependent on desorption of methane.
Surface properties The theory presented for charcoal and benzine (Gregg, 1961) appears to be true for the coal and methane system as well. Figure 12a shows a molecule of methane adsorbed on the surface of solid coal - a micropore within the matrix. Assuming a purely physical adsorption, a part of the surface energy of coal is utilized in retaining the methane molecule on the site shown. Hence there is an overall reduction in the free surface energy of coal. Figure 12b shows the same site after the methane molecule is desorbed. The energy that was being
Harpalani and Schraufnagel
378
o-=~=,,-egauge
Exhaust-~E~'----
~
/
~Gas
~-~E----~
strain
To indicator
From s t r a i n -
gauges~2
ecimen
Axial strain'gauge#: 1
dial strain ~1
uge
777 Fig. 8. Experimental set-up to monitor volumetric changes in coal matrix with changing gas pressure
used in keeping the methane molecule is no longer required. The free surface energy of the adsorbent - coal - is thus increased. This energy now acts as a force inwards, i.e. it pulls the surface inwards. This inward pull would cause the adsorbent to shrink. If this shrinkage takes place throughout the matrix - in between the cleats on all sides - the entire matrix block would shrink inwards. This would increase the aperture of the cleats (see Fig. 2) thus resulting in increased permeability and increased gas flow.
Impact of stress The situation in situ is more complicated due to stess and presence of water in the cleats,but the permeability would nevertheless increase. In fact, permeability variation - as a function of pressure - does indicate an increase wheo accompanied by desorption of methane. This was observed during the gas pressure-permeability experiments, using cylindrical specimens 1.5 in. long under triaxial stress conditions. The results have been published by Harpalani (1989). The general relationship is shown in Fig. 13 for decreasing gas pressure-desorption. It clearly indicates that once significant desorption starts, permeability starts to increase and continues to rise dramatically.
Measurement of parameters impacting methane recovery from coal seams
#
Experiment
~
~
/
.=:
3
"--
2
OI
~
I
379
/
I
E x p e r i m e n t ~t:l •
A d s o r p t i "o n
I
I
;
1 O0 Helium cycle
-1
Gas pressure (psi)
Fig. 9. Volumetric strain of coal matrix for two cycles: helium and methane 6 E f f e c t i v e shrinkage
5
Methane
4
3 ._ o
2
>
0
f 200
= 400
I 600
I 800
~
-1 Gas
I 1000 Helium
pressure
(psi)
Fig. 10. Volumetric strain of coal matrix for decreasing gas pressure and the effective shrinkage due to desorption of methane The current understanding of the depth-permeability relationship suggests a rapid decline in coal reservoir permeability with increasing depth (Koenig, 1989). However, drilling and testing in the deeper coals in the basins reveal much higher permeabilities than those predicted by the 'existing theory' shown in Fig. 14. The relationship between depth and
Harpalani and Schraufnagel
380
Effective
6
shrinkage
28 5
~
24
) 20
c
g
o
8
16
1 4
o
0
i
I
I
200
400
600
I
800
Gas pressure (psi)
Fig. 11. Adsorbed gas and effective shrinkage with decreasing gas pressure
•
Meth;ane molecule (adsorbed)
Coal
/~
Co~l
Fig. 12. Changes in free surface energy of coal with adsorption/desorption
I
1000
Measurement of parameters impacting methane recovery from coal seams m Adsorbed
mesh)
• Perrneab~llty (hydrostat;c stress : 1 5 0 0 psi )
6
5
gas (28-48
381
25
20 co
3
15
1
10
i \
j.
5
0
T
I 0
Gas
I 800
!
40q
200
600
pressure
I
1000
(psi)
Fig. 13. Variation in adsorbed gas and permeability of a coal sample for decreasing gas pressure Positive e f f e c t of m a t r i x shrinkage
100
}
,o
g "~
50
-~ ~. (.9
Negative e f f e c t of of pore Volume compressibility 30
10
i
i
I
w
i
I
2
3
4
.5
Time (years)
Fig. 14. Seam permeability against depth (theory and observation)
382
Harpalani and Schraufnagel
permeability is greatly influenced by localized in situ stresses and permeability enhancement. Figure 14 shows that permeabilities two orders of magnitude higher than previously expected for deep coals are indeed possible (Schraufnagel et al., 1990). Overall sensitivity to production To evaluate the sensitivity of gas production to coal matrix shrinkage, the experimental results were used as input in a set of simulation runs using a coalbed methane simulator, COMETPC 3-D model, developed by ICF Resources (Harpalani, 1989; Sawyer et al., 1990). The effect of pore volume and grain compressibilities was studied on 5-year gas production. A general relationship derived from this analysis is shown in Fig. 15. Case 1 is the basis of comparison, with zero pore volume compressibility and no matrix shrinkage. Case 2 shows a 15% reduction in gas production due to the negative effect of pore volume compressibility (Cp=4.5 x 10 -s per psi).
fl 11 E i,.
~
E nh ance pe rm edaytb ili
/ I
I
I
I
Depth
Fig. 15. Effectof the two compressibilities on gas production Case 3 shows that matrix shrinkage can more than offset this reduction. In fact, there is an increase of approximately 60% as a result of matrix shrinkage compressibility (6.2 x 10 .6 per psi). Coal matrix shrinkage is therefore an important parameter and must be considered together with cleat contraction due to pore volume compressibility in order to simulate accurately and forecast long-term gas production from coalbed methane wells. These initial findings on coal permeability are encouraging for the development of coalbed methane, particularly for deeper and tighter coals. Gas production from deep coalbeds has been limited due to the high in situ stresses. Although the depth-permeability relationship is
Measurement of parameters impacting methane recovery from coal seams
383
complex, should these new data prove to be representative, the commercial potential of coalbed methane would increase substantially.
Conclusions The most important finding of this study is that matrix shrinkage can potentially impact long-term gas production from coalbed methane reservoirs. Desorption of methane results in shrinkage of coal matrix. The shrinkage of matrix, in turn, increases the permeability due to expansion of the cleats. The increase in permeability due to matrix shrinkage is more than the decrease due to cleat contraction resalting from pressure depletion, and there is an overall increased permeability. This increase in permeability can result in substantial increased production over the life of methane-producing wells. It may also make methane recovery from deep reservoirs economically feasible due to enhanced permeability- particularly during the later part of the productive life of a well. On the more general side, the microscopy study confirms the usage of a dual-porosity model for stimulation purposes. The micrographs show the two - fracture and micropore porosities. The results also confirm the large area of the internal surface of coal explaining how such large quantities of methane are stored within it. Although the internal surface area is seen to be very large, the porosity of the samples was determined experimentally to be very low - approximately 5%. This suggests that a good number of the pores are either of the 'dead end' type, or disc-shaped with large surface area but very small volume. Based on the work described in this paper, it is felt that the following areas should be investigated further: 1. More measuresments of permeability and matrix shrinkage on other types of coal should be made to confirm the effect for all coals. 2. Effect of stress on matrix shrinkage should be investigated in detail. This phenomenon could be particularly significant for deep coalbeds- like the Piceance basin-where extremely low permeabilities have been encountered. The effect of the presence of water in cleats on matrix shrinkage should be investigated as well. This would give results of the combined effect of matrix shrinkage and relative permeability of coal to gas and water. 3. Pore volume compressibility (Cp) and matrix shrinkage compressibility (Cm) should be established for US coals. This should be incorporated in the simulation models used commonly. Sensitivity of permeability and production to compressibility should be investigated in detail.
Acknowledgement This research was carried out as a part of a research project sponsored by the Gas Research Institute, Contract No. 5088-215-1666. The authors thank the GRI for support.
References Cervik, J. (1967) The behavior of coal-gas reservoirs. Paper presented at the SPE-AIME Eastern Regional Meeting, Pittsburgh, PA (SPE 1973).
384
Harpalani and Schraufnagel
Dabbous, M.K. et al. (1974) The permeability of coal to gas and water. SPE Journal, 563-72. Gregg, S.J. (1961) The Surface Chemistry of Solids, Reinhold Publishing Corp., New York, pp. 42-3. Harpalani, S. (1989) Permeability changes resulting from gas desorption, Final report submitted to Gas Research Institute, Contract No. 5088-215-1666. Harpalani, S. and McPherson, M.J. (1985) Effect of stress on permeability of coal, Quarterly Review of Methane from Coal Seams Technology, 3, 23-8. Harpalani, S. and McPherson, M.J. (1986) Retention and release of methane in underground coal workings, International Journal of Minin9 and Geological Engineering, 4, 217-33. Jones, A.H. and Bell G.J. (1987) The influence of coal fines on the behavior of hydraulic fracture stimulation treatments, in Proceedings of the Coalbed Methane Symposium, University of Alabama, Tuscaloosa, pp. 93-102. Kamal, M.M. and Six, J.L. (1989) Pressure transient testing of methane producing coalbeds, presented at the 64th Annual Technical Conference of the Society of Petroleum Engineers, San Antonio, TX (SPE 19789). King, G.R. (1985) Numerical simultaion of the simultaneous flow of methane and water through dual porosity coal seams during the degasification process, Ph.D. thesis, Pennsylvania State University. Koenig, R.A. et al. (1989) Application of hydrology to evaluation of coalbed methane reservoirs, Final report submitted to Gas Research Institute, Contract No. 5087-214-1489. Reznik, A.A. et al. (1974) Air-water relative permeability studies of Pittsburgh and Pocahontas coals, SPE Journal, 566-72. Sawyer, W.K. et al. (1987). Using reservoir simulation and field data to define mechanisms controlling coalbed methane production. Proceedings of the Coalbed Methane Symposium, University of Alabama, Tuscaloosa, 295-307. Sawyer, W.K. et al. (1990) Development and application of a 3D coalbed simulator, presented at CIM/SPE 12th Technical Meeting, Calgary (CIM/SPE 90-119). Schraufnagel, R.A. et al. (1990), Coalbed methane development faces technology gaps, Oil and Gas Journal, February 5, 48-54. Somerton, W.H. et al. (1975), Effect of stress on permeability of coal, Int. J. of Rock Mech. Min. Sci. and Geo. Abstr. 12, 129-45. Taber, J.J. et al. (1974) Development of techniques and the measurement of relative permeability and capillary pressure relationships in coal, Final Report prepared for US Bureau of Mines, Contract No. G0122006 (NTIS No. PB-232244). Zuber, M.D. et al. (1987). The use of simulation and history matching to determine critical coalbed methane reservoir properties, presented at the Low Permeability Reservoirs Symposium, Denver (SPE/DOE 16420).