Several investigators have reported coal permeability decreases with increasing stress, but no conceptual model has been advanced to explain this effect. To better understand the permeability of stressed coal, a theoretical and experimental program was undertaken. A common naturally fractured reservoir geometry, a collection of matchsticks, was extended to stressed coalbeds and tested against laboratory measurements using samples from the San Juan and Warrior Basins. Good agreement was obtained between theoretical behavior and laboratory data. Equations are presented for converting laboratory measured stress-permeability data to (a) in-situ permeability as a function of depth of burial in a basin, and (b) to reservoir permeability during coalbed depletion.Coal cleat compressibility, analogous to pore volume compressibility of conventional reservoirs, has historically been difficult and expensive to measure and the results of such measurements are often ambiguous. A method is presented for calculating cleat volume compressibility from stress permeability experiments, resulting in considerable savings of both time and money. Stress-permeability and cleat volume compressibility results reported here are compared with those published in the literature.Evidence in the literature indicates that coal matrix shrinks when gas is desorbed, increasing cleat permeability. Assuming a matchstick geometry and using a coal matrix shrinkage coefficient reported in the literature, the increase in cleat permeability due to matrix shrinkage was calculated. The increase in permeability due to matrix shrinkage during depletion is compared with the decrease in permeability due to increased stress.References and illustrations at end of paper.

Introduction Deliverability of coal wells, like conventional gas wells, depends on bottomhole flowing pressure. Because coal wells often produce both gas and water, lowering bottomhole flowing pressure to increase gas rate also increases water rate. Thus, optimization of coal well profitability entails balancing gas revenues and water disposal costs. The present study was undertaken to determine if the relation between coal well bottomhole flowing pressure and gas and water production rates could be described by Vogel's Inflow Performance Relation (IPR). First, simulation studies were clone to test the applicability of Vogel's IPR to coal wells. Secondly, productivity of actual coal wells was compared with Vogel's IPR curves. Productivity of an oil well draining a solution-gas drive reservoir was investigated by Vogel using numerical simulation. A total of 21 simulations covering a wide range of oil, PVT properties, and relative permeabilities were made. By using dimensionless pressures and rates, Vogel found well productivity could be described by (1) where q is oil production rate in bpd, qmax is maximum oil production rate in bpd, pwf is bottomhole flowing pressure in psia, and pavg is average reservoir pressure in psia. Eq. (1), called Vogel's Inflow Performance Relation (IPR), was found to describe Simulated well productivity with a typical accuracy of 10%. Errors as high as 20% were noted for simulations of viscous crudes and/or damaged wells with skin factors great than +5. Over the last quarter century, Vogel's IPR curve has been extensively used to pi-edict oil well performance. Because of his success, the question arose as to whether Vogel's IPR could also describe gas and water production from a coal well. A more generalized Inflow Performance Relation was developed by Richardson and Shaw. Their equation (2) includes a variable Vogel coefficient, denoted as V. The Vogel coefficient, corresponding to the classic Vogel IPR is 0.2. Because of the unconventional nature of coalbed methane, it was speculated that the generalized method of Richardson and Shaw may describe coal well gas and water production more accurately than the Vogel IPR. Deliverability of a Warrior Basin coal well was discussed by Reeves et al. Backpressure of a Deerlick Creek well was sequentially decreased then increased, with the well being kept at each pressure for a week. Data obtained from the increasing bottomhole pressure steps fell substantially below that obtained from the decreasing bottomhole pressure steps. Reeves et al. attributed this behavior to the well not being stabilized. As shown by Mavor and Robinson, stabilization times of coal wells can be much longer than conventional gas wells with similar permeabilities and pressures clue to sorption compressibility. In the present study, care was taken in the field work to ensure wells were stabilized at a given pressure before moving to the next point. Simulation Studies Due to the expense of field tests, simulation was first used to investigate applicability of Vogel's IPR to coal wells. This study used Amoco's fully implicit, conventional reservoir simulator modified for coalbed methane simulation as described by Seidle and Arri. P. 641^

No abstract