A Parametric Study on the Benefits ofDrilling Horizontal and Multilateral Wells in

Coalbed Methane ReservoirsNikola Maricic, SPE, Chevron Corporation; Shahab D. Mohaghegh, SPE, West Virginia University, and

Emre Artun, SPE, Chevron Corporation

SummaryRecent years have witnessed a renewed interest in development ofcoalbed methane (CBM) reservoirs. Optimizing CBM productionis of interest to many operators. Drilling horizontal and multilateralwells is gaining popularity in many different coalbed reservoirs, withvarying results. This study concentrates on variations of horizontal-and multilateral-well configurations and their potential benefits.

In this study, horizontal and several multilateral drilling pat-terns for CBM reservoirs are studied. The reservoir parameters thathave been studied include gas content, permeability, and desorp-tion characteristics. Net present value (NPV) has been used as theyard stick for comparing different drilling configurations. Configu-rations that have been investigated are single-, dual-, tri-, andquad-lateral wells along with fishbone (also known as pinnate)wells. In these configurations, the total length of horizontal wellsand the spacing between laterals (SBL) have been studied. It wasdetermined that in the cases that have been studied in this paper(all other circumstances being equal), quadlateral wells are theoptimum well configuration.

IntroductionSeveral studies in the past have examined the impact of differentreservoir properties on CBM production (Ertekin et al. 1988;Zuber 1998; Katyal et al. 2007), including the effect of secondary(micro-) permeability (Gamson et al. 1993; Mavor and Gunter2006; Palmer et al. 2006). This paper is dedicated to examining theimpact of different horizontal-well configurations (multilateral) onthe production and recovery in CBM reservoirs.

In this section, a brief background on reservoir-engineeringaspects of CBM production is presented along with details on thereservoir simulator used for modeling the cases used in this study.Some basic introduction about horizontal wells and their impact onproduction behavior is also presented.

CBM Background. Coal represents an unusual reservoir rock,resulting from its highly complex reservoir characteristics. One ofthe characteristics that distinguish coal seams from conventionalgas reservoirs is that coal represents both the source and the res-ervoir rock at the same time. Unlike conventional sandstone res-ervoirs, where the gas is found in free state within the pore struc-ture of the rock, the methane gas is adsorbed onto the internalstructure of the coal, which allows a significant amount of gas tobe stored in the coal rock (Saulsberry et al. 1996). The coal-seamsystem is a complex heterogeneous reservoir characterized by anaturally fractured system, macropores (also known as the cleatsystem), and micropores (the coal-matrix system). In general, thecoal cleat system is orthogonal, with one direction cross cutting theother, and it varies from case to the case with significant impact onthe coal deliverability (Remner et al. 1986). The CBM productiondepends highly on the fracture system, fracture spacing, and frac-ture interconnection. If the cleat system for any reason is not

developed enough, gas production could be difficult. This occursbecause of the low values of porosity and permeability in thematrix, making it almost impossible for gas to move from thematrix into the fractures. At the beginning, the coal system is inequilibrium, and typically, water must be produced continuouslyfrom coal seams to reduce the reservoir pressure and release thegas. Gas from the coal can be produced only after initial dewater-ing of the system and upon reaching low reservoir pressure. Oncethe pressure drops lower than the “critical desorption point,” meth-ane desorbs from the coal surface (Guo et al. 2003). The dewa-tering process can take anywhere from a few days to severalmonths, which depends on CBM-well configuration. Generally,the water production declines until the gas rate reaches the peakvalue. This “time to peak gas” is a critical parameter because thegas production starts declining after reaching the maximum(Roadifer et al. 2003). Upon reaching the peak, gas productionstarts to decline and behavior of CBM production becomes similarto that of conventional reservoirs.

Simulation Details. All simulation models for this study wereperformed by use of a commonly used commercial simulator[CMG’s Advanced Compositional Reservoir Simulator (GEM2006)]. The commercial simulator used in this study is a fullycompositional model used to model reservoir-engineering caseswhere fluid composition and their interactions are essential to theunderstanding of the recovery process. It is an engineering tool formodeling complex reservoirs with complicated phase-behavior in-teractions, which impact directly on the recovery mechanisms usedto optimize the recovery (Arenas 2004).

In our implementation of the simulation runs, automatictimestep was used in the range of 0.01 to 365 days. Furthermore,an adaptive-implicit solution method (fully implicit strategy) withstability switching was used where wellblocks and neighboringblocks are set implicitly. All other blocks are checked in everytimestep for stability after convergence (GEM 2006). The simula-tion runs represented a dual-porosity model with the matrix-to-fracture and fracture-to-fracture flow calculations.

A formulation of the Gilman and Kazemi type was used for theshape-factor calculation and a pseudocapillary pressure modelwith corrections was used for the matrix-to-fracture transfer cal-culations. Rock-compressibility values used in the study at thestate of equilibrium were 1.0 E−06 psi−1 at reference pressure of412 psi for the matrix and fractured systems. The Peng-Robinsonequation of state was selected, and the constant reservoir tempera-ture was set as 113°F. Relative permeability curves shown in Fig. 1were used for all the simulation runs that are presented in thispaper, with an initial water saturation of 92.7% in the cleat system.

It must be noted that during this study, matrix shrinkage owingto production and (therefore) its potential impact on stress-relatedpermeability were not taken into account. Although such consid-erations may have impacted the overall production behavior, itwould not have impacted the overall conclusions of this study.Readers are encouraged to see Shi and Durucan (2005) and Gu andChalaturnyk (2005) for details on this topic.

Horizontal Wells. Methane can be drained through small-diameter horizontal holes. The main advantages of horizontal wellsare that the well direction, shape, and position can be controlled.

Copyright © 2008 Society of Petroleum Engineers

This paper (SPE 96018) was accepted for presentation at 2005 SPE Annual TechnicalConference and Exhibition, Dallas, 9–12 October, and revised for publication. Originalmanuscript received for review 11 July 2005. Revised manuscript received for review 10July 2008. Paper peer approved 2 August 2008.

976 December 2008 SPE Reservoir Evaluation & Engineering

By use of horizontal wells, an almost ideal position of the well canbe determined and implemented with respect to principle perme-ability directions of the coal (Osisanya and Schaffitzel 1996).Proper positioning of the well, and proper length of the borehole,can contribute significantly to draining large areas.

An effective stimulation technique may be a horizontal bore-hole placed perpendicular to face-cleat direction, thus providingmaximum access to the primary flow channels (Logan 1988). Onequestion is which shape of horizontal well to drill to produce largeramount of gas in the shortest possible time. Drilling horizontalwells in coal seams is highly important to the sweep efficiency. Asthe length of the well and contact of wellbore to coal seam in-crease, the time for water production decreases. Usually, the gas-production curves of vertical and horizontal wells will differ sig-nificantly from each other. In a short time, the horizontal well willdewater the system, and significant water production will takeplace. The gas-flow peak will occur soon after the well starts toproduce. The most important part of the gas-flow curve is probablyjust after gas flow peaks. The slope of the curve is importantbecause the well will produce like a conventional gas reservoirfrom that point. The flatter the curve, the better production will befor the rest of the well’s life. The issue is that horizontal wells arecapable of draining the system fast because of much longer well-bore contact with the reservoir. Consequently, the adsorption pro-cess is triggered quickly, and the well will elicit a steep negativedecline curve relatively fast. After reaching the peak, the gas pro-duction will start to decrease, again owing to the large contact areabetween the horizontal well and the reservoir. Vertical wells, al-

lowing an exceedingly small contact between them and the coalseam, are rarely, if ever, acceptable producers. Production rates donot exceed a few thousand cf/D, and they are hindered further byrelative permability problems caused by the presence of associatedwater (Deimbacher et al. 1992). Figs. 2 and 3 demonstrate compari-son of production rates and of cumulative production, respectively,between vertical and horizontal wells from a typical CBM reservoir.

Parametric Study: Systematic ApproachIn this study, a number of simulations have been performed toinvestigate the influence of different horizontal-well shapes, theirvarious length, SBL, and various reservoir properties on the de-liverability of a certain type of coal. To explain various stepsperformed in this study better, a flow chart is used (Fig. 4). Thestarting point of this research was determining the different hori-zontal-well configurations that would be used and implemented inthe study. As a result, five horizontalwell shapes were examined:single-lateral, dual-lateral, trilateral, quadlateral, and fishbone(pinnate). No matter which horizontal shape has been applied tothe reservoir simulator, the vertical length (depth from the surfaceto the horizontal-kickoff point) is constant in all cases during thisstudy. Definitions of a lateral well, SBL, and vertical depth areillustrated in Fig. 5. Investigated well configurations are illustratedin Fig. 6.

Sensitivity analysis (parametric study) represents a set of simu-lation runs in which all parameters are kept constant, but only oneis changed systematically. This approach enables us to monitorchanges of simulation outputs influenced by only one variable. Thefirst step taken in this part of the approach was to develop a modelrepresenting CBM and test it by use of different well configura-tions. A dual-porosity model was used to create a CBM reservoir

Fig. 1—Relative permeability curves used in this study.

Fig. 2—CBM-production rates from horizontal and vertical wells.

Fig. 3—CBM cumulative production from horizontal and verti-cal wells. Fig. 4—Flow chart of the study presented here.

977December 2008 SPE Reservoir Evaluation & Engineering

including a Gilman-Kazemi shape-factor calculation. A Cartesian-grid model has been chosen for the modeling of the CBM reser-voir, which has the input parameters shown in Table 1. This ho-mogeneous model by use of a Cartesian grid has been formed witha constant drainage area of 320 acres, consisting of 30×30 squareblocks with each side measuring 120 ft. As previously mentionedin this section, regardless of the horizontal shape applied to thereservoir simulator, the vertical length is constant in all casesduring this study.

The production period has been determined as 15 years. Theparametric study is divided into three parts. In the first part, theinfluence of well configuration and SBL is investigated consider-ing the five configurations mentioned earlier in this section. In thesecond part, the influence of several reservoir/coal properties suchas permeability, time constant, gas content, and desorption char-acteristics is investigated. The third part involves a study on thick-ness. In each part, gas recovery and NPV results at the end of the15-year production period are presented. NPV is calculated at adiscount rate of 8% after considering the costs associated withdrilling operations. The most important cost factor was the hori-zontal drilling cost, which was assumed as USD 100/ft after dis-cussion with industry professionals. It should be noted that this isan average value for a drilling operation under perfect conditions,and it might increase as problems arise during drilling. Futher-more, remedial operations or standard operating procedures suchas wellbore cleanups are not considered in the calculation of NPV.Such operations can increase well operating costs significantly anddrive down NPV.

Well Configuration/SBL. The starting point in this part of thestudy is the vertical well placed in the middle of the reservoir withhorizontals spreading out toward the upper-left corner of the model(Fig. 7). The single-lateral well configuration was the first one tobe investigated. By definition, this was the only case in the studywhere SBL was not used.

Well configurations with different horizontal lengths weresimulated. The drainage area was kept constant, and for the wholeset of simulations, only the horizontal length of the well and SBLwere continuously changed step by step. Once SBL was deter-mined (the narrowest one), it remained constant, changing onlyhorizontal length from the shortest to the longest length whilerunning the simulator for each different scenario that took place.The maximum length of a horizontal well is constricted by theshape and the size of the reservoir. With dual-, tri-, and quad-laterals, SBL was introduced as a variable. The process describedfor single-lateral wells was applied to the rest of investigated wellshapes, taking into account the new variable (i.e., SBL). Uponcreating all possible scenarios for a minimum value of SBL, a newSBL with a higher value than the previous one was introduced. Again,the same procedure took place, keeping SBL constant while changinghorizontal length from minimum to maximum. After the completionof all possible cases for one configuration of horizontal wells, anotherconfiguration was introduced. The same procedure was applied totri-lateral and quad-lateral horizontal-well configurations.

The problem with the pinnate shape required some modifica-tion of this approach. Obviously, there is no SBL with the single-lateral horizontal configuration, but dual-, tri-, and quad-lateralconfigurations possess SBL, which was investigated in these three

cases. For the pinnate horizontal-well shape, another scenario wasneeded because of the specific shape of this well configuration.Another variable was introduced: namely the number of laterals.

Reservoir Properties. The second part of the sensitivity analysisrequired the definition of coalbed-reservoir properties subject tochange during the study. Knowing all the steps performed in thefirst part of this analysis, this time the minimum well length andSBL were defined and kept constant, and the reservoir data werechanged one by one. The goal of this approach was to introduce anew dimension to the model, which is changing one reservoirproperty at a time. The quad-lateral well configuration with 680 ft

Fig. 6—Investigated well configurations.

Fig. 5—Definitions of terms related to a lateral well.

978 December 2008 SPE Reservoir Evaluation & Engineering

of SBL has been used to conduct parametric studies to assess theeffects of parameters such as permeability, time constant, and gascontent. That configuration was chosen because it was the one thathas yielded the best gas-recovery results.

Thickness. In this part of the study, a relatively smaller model hasbeen developed to investigate gas recovery in different cases onthe basis of pay thickness. A 15×15-gridblock model was devel-oped with the same gridblock size of 120 ft. Vertical depth waskept constant. Cases of a single layer with net-pay thicknesses of4 ft and 12 ft and a case of three layers with a thickness of 4 ft eachwere simulated with well configurations of single-vertical, four-spot, and horizontal wells.

Results and DiscussionWell Configuration/SBL. With values of the actual gas produc-tion as the simulator output and calculating initial gas in place as820 MMscf and initial water in place as 520,000 bbl, gas recoveryfor each well configuration has been calculated and plotted againsttotal horizontal length of the well. Gas recovery vs. total-horizontal-length shape behavior is shown in Fig. 8. Gas-recovery

Fig. 7—Single horizontal-lateral wells with different horizontal lengths.

Fig. 8—Gas recovery for single-lateral well configurations.

979December 2008 SPE Reservoir Evaluation & Engineering

results for dual-, tri-, and quad-lateral and pinnate configurationsare shown in Figs. 9 through 12, respectively.

The best producers from five different well configurations havebeen compared with one another (Fig. 13). Our investigationshows that the quadlateral well configuration with an SBL of 680ft and total horizontal length of 8,000 ft is the one with the highestgas recovery (approximately 36%). The strongest competitor, thepinnate well configuration, has a little better recovery (38%) butfor the total horizontal length of 18,000 ft. But the question iswhether to drill an additional 10,000 ft of horizontal section torecover 2 to 3% more gas. If these two horizontal-well configu-rations are compared for the same total horizontal length of 8,000ft, it is obvious that the quadlateral well configuration can recover5 to 10% more gas.

Although a larger horizontal length increases the total gas re-covery, increasing horizontal-drilling costs must be considered tofind out the optimum configuration. Fig. 14 shows NPVs for thequad-lateral well configuration, which is now the recommendedwell configuration. The graph shows the results with respect to thehorizontal length and SBL. This graph suggests the optimum con-figuration would be the quad-lateral well configuration with anSBL of 680 ft and a total horizontal length of 3,100 ft.

Reservoir Properties. The quad-lateral well configuration with680 ft of SBL and with a horizontal length of 8,000 ft has beenused to investigate the influence of parameters such as permeabil-ity, time constant, and gas content.

Permeability. The fracture permeability in i and j directions hasbeen changed together so as to have values of 2, 5, 8, 10, and 15md, while the permeability in the k direction is kept constant at 2md. Gas recovery after 15 years of production has increased withincreasing permeability, as expected. Permeability values chang-ing between 2 and 15 md had a corresponding range of 0.19 and

0.35 of overall gas recovery (Fig. 15). Fig. 16 reflects the NPVvalues for each case, and it shows a similar trend.

Time Constant. Values of 10, 50, 150, and 200 days have beenassigned to the time constant, � (i.e., coal-desorption time), to seethe changes in the reservoir behavior resulting from the time-constant parameter. After 15 years of production, overall gas re-covery did not differ too much with varying values of the timeconstant. It has been observed that overall gas recovery wouldhave a range of 0.26 and 0.29 for time-constant values changingbetween 10 days and 200 days (Fig. 17).

Because it is a known fact that the time constant mainly affectsthe earlier production from a CBM reservoir, it has been decidedto consider the first year’s production in each case. Fig. 18 showsthe cumulative gas production during the 15 years of productionfor each case with different time-constant values ranging from 10to 200 days.

As discussed, as far as the long-term production (i.e., 15 yearsin this case) is concerned, time constant does not have a significanteffect on overall gas recovery. While the gas produced after thefirst year is approximately 120 MMcf for a time constant of 10days, when the time constant is as large as 200 days, the first year’sproduction is approximately 80 MMcf. Considering the expecteddifference in production of 40 MMcf of gas, it can be concludedthat the time constant has a significant effect on the first year’sproduction of the reservoir. It is observed that the difference incumulative production after the first year remains constant during15 years of production. Assuming a gas price of USD 5/Mscf, thisdifference would be equal to as much as USD 200,000. This dif-ference is clearly reflected in Fig. 19, which shows NPV values foreach case.

Gas Content. Gas content is one of the most important param-eters that influences the gas recovery from coalbeds.

Fig. 9—Gas recovery for dual-lateral well configurations. Fig. 10—Gas recovery for tri-lateral well configurations.

Fig. 11—Gas recovery for quad-lateral well configurations. Fig. 12—Gas recovery for pinnate well configurations.

980 December 2008 SPE Reservoir Evaluation & Engineering

Gas content is a measurement of the actual gas contained in agiven resevoir (Schraufnagel and McBane 1994). The same systemhas been simulated with different values of gas content as 350,395, and 450 scf/ton. Considering these values, five different iso-therms have been generated by changing Langmuir volume andpressure constants, which are controlled by maximal adsorbedmass and Langmuir adsorption constant, respectively. Character-istics of these five isotherms are shown in Table 2.

One of the isotherms is an average isotherm for Northern Ap-palachia, and others are defined as modifications to that isotherm.Initial reservoir pressure is kept at 450 psi for all cases. The graphthat shows the behavior of each isotherm is shown in Fig. 20,which indicated the corresponding gas content with respect topressure for the saturated case for each isotherm. The final gasrecoveries after 15 years of production for each isotherm areshown in Fig. 21 and the corresponding NPV are shown in Fig. 22.The isotherms are ranked on the basis of the recovery as Modifi-cation 3, Northern Appalachia, Modification 1, Modification 2,and Modification 4, respectively. Considering the isotherm curvesshown in Fig. 20, the decline behaviors of the five isothermssupport the gas-recovery values. Fig. 21 shows the NPV results foreach case, which show a trend similar to the one obtained for gasrecovery.

Thickness. Fig. 23 shows gas-recovery results for each well con-figuration and for each case, which include a case with a singlelayer of 4-ft thickness, a case with a single layer of 12-ft thickness,and the case with three layers, each with a thickness of 4 ft. Thisgraph clearly shows that horizontal-well configurations bring thebest gas recovery. The large gap between the gas-recovery resultsfor the case with a single layer of 4 ft and others is caused by thedifference in the initial gas in place. In our analysis, the case withthe single layer could not recover its investment. The case withthree layers had an NPV profile like the one shown in Fig. 24. This

graph also clearly shows that drilling horizontal wells is morefavorable, than drilling vertical wells.

ConclusionsA parametric study to determine the benefits of drilling horizontaland multilateral wells in CBM reservoirs is presented. Various wellconfigurations, such as single-, dual-, tri-, and quad-lateral and fish-bone (also known as pinnate), were investigated while consideringparameters such as SBL and total horizontal length. Sensitivity analy-ses also were conducted to examine the effects of thickness, numberof layers, and various reservoir/coal parameters such as permeabil-ity, time constant, and gas content. It was noted that gas content isa major contributor to gas recovery from CBM reservoirs. Majorconclusions of this study can be summarized as follows:1) The optimum well configuration can be determined by consid-

ering the total horizontal length, SBL, and number of laterals.Higher horizontal length increases the contact with the coalseam and yields a higher gas recovery, but it also increases theassociated drilling costs at the same time. After considering theeconomics, the optimum configuration, which is suggested bythis study, is a quad-lateral well configuration with SBL of 680ft and total horizontal length of approximately 3,100 ft.

2) Time constant has no significant effect on the overall recovery.It is an important factor that influences the production in thefirst year, however, and thus the economics of the first yearof production.

Nomenclaturei, j, k � indecies identifying directions x, y, z, respectively

kr � relative permeability, fractionkrw � water relative permeability, fractionkrg � gas relative permeability, fractionSw � water aturation, fraction

Fig. 13—Gas recovery comparing best producers. Fig. 14—NPV of quad-lateral horizontal-well configurations.

Fig. 15—Effect of permeability on gas recovery. Fig. 16—Effect of permeability on NPV.

981December 2008 SPE Reservoir Evaluation & Engineering

PL � Langmuir pressure constant, m/Lt2, psiVL � Langmuir volume constant, L3, scf

� � desorption time constant, t, days

AcknowledgmentThe authors would like to thank Computer Modelling Group Ltd.for providing the software used in this study.

ReferencesArenas, A.G. 2004. Development of gas production curves for coalbed

methane reservoirs. MS thesis, West Virginia University, Morgantown,West Virginia.

Deimbacher, F.X., Economides, M.J., Heinemann, Z.E., and Brown, J.E.1992. Comparison of Methane Production From Coalbeds Using Ver-tical or Horizontal Fractured Wells. JPT 44 (8): 930–935; Trans.,AIME, 293. SPE-21280-PA. DOI: 10.2118/21280-PA.

Ertekin, T., Sung, W., and Schwerer, F. 1988. Production PerformanceAnalysis of Horizontal Drainage Wells for the Degasification of CoalSeams. JPT 40 (5): 625–632. SPE-15453-PA. DOI: 10.2118/15453-PA.

Gamson, P.D., Beamish, B.B., and Johnson, D.P. 1993. Coal Microstruc-ture and Micropermeability and Their Effects On Natural-Gas Recov-ery. Fuel 72 (1): 87–99. DOI: 10.1016/0016-2361(93)90381-B.

GEM Advanced Compositional Reservoir Simulator, Version 2006 UserGuide. 2006. Calgary, Alberta: CMG.

Gu, F. and Chalaturnyk, R.J. 2005. Analysis of coalbed methane produc-tion by reservoir and geomechanical coupling simulation. J. Cdn. Pet.Tech. 44 (10): 33–42.

Guo, X., Du, Z., and Li, S. 2003. Computer Modeling and Simulation ofCoalbed Methane Reservoir. Paper SPE 84815 presented at the SPEEastern Regional Meeting, Pittsburgh, Pennsylvania, 6–10 September.DOI: 10.2118/84815-MS.

Katyal, S., Valix, M., Thambimuthu, K. 2007. Study of parameters affect-

Fig. 17—Effect of time constant on gas recovery.

Fig. 18—Effect of time constant on cumulative gas production.

Fig. 19—Effect of time constant on NPV.

Fig. 20—Five isotherms defined with different pressure and vol-ume constants (PL and VL).

Fig. 21—Final gas recoveries for five isotherms after 15 yearsof production.

982 December 2008 SPE Reservoir Evaluation & Engineering

ing enhanced coal bed methane. Energy Sources, Part A: RecoveryUtilization and Environmental Effects 29 (3): 193–205.

Logan, T.L. 1988. Horizontal Drainhole Drilling Techniques Used for CoalSeam Resources Exploitation. Paper SPE 18254 presented at the SPEAnnual Technical Conference and Exhibition, Houston, 2–5 October.SPE-18254-MS. DOI: 10.2118/18254-MS.

Mavor, M.J. and Gunter, W.D. 2006. Secondary Porosity and Permeabilityof Coal vs. Gas Composition and Pressure. SPEREE 9 (2): 114–125.SPE-90255-PA. DOI: 10.2118/90255-PA.

Osisanya, S.O. and Schaffitzel, R.F. 1996. A Review of Horizontal Drillingand Completion Techniques for Recovery of Coalbed Methane. PaperSPE 37131 presented at the International Conference on HorizontalWell Technology, Calgary, 18–20 November. DOI: 10.2118/37131-MS.

Palmer, I.D., Cameron, J.R., and Moschovidis, Z.A. 2006. Permeabilitychanges affect CBM production predictions. Oil & Gas J. 104 (28):43–50.

Remner, D., Ertekin, T., Sung, W., and King, G. 1986. A Parametric Studyof the Effects of Coal Seam Properties on Gas Drainage Efficiency.SPERE 1 (6): 633–646. SPE-13366-PA. DOI: 10.2118/13366-PA.

Roadifer, R.D., Moore, T.R., Raterman, K.T., Farnan, R.A., and Crabtree,B.J. 2003. Coalbed Methane Parametric Study: What’s Really Impor-tant to Production and When? Paper SPE 84425 presented at the SPEAnnual Technical Conference and Exhibition, Denver, 5–8 October.DOI: 10.2118/84425-MS.

Saulsberry, J.L., Schafer, P.S., and Scraufnagel, R.A. eds. 1996. A Guideto Coalbed Methane Reservoir Engineering. Technical Report GRI-94/0397, Gas Research Institute, Chicago, Illinois.

Schraufnagel, R. and McBane, R. 1994. Coalbed Methane—A Decade ofSuccess. Paper SPE 28581 presented at the SPE Annual TechnicalConference and Exhibition, New Orleans, 25–28 September. DOI:10.2118/28581-MS.

Shi, J.-Q. and Durucan, S. 2005. A Model for Changes in Coalbed Per-meability During Primary and Enhanced Methane Recovery. SPEREE8 (4): 291–299. SPE-87230-PA. DOI: 10.2118/87230-PA.

SI Metric Conversion Factorsacre × 4.046 873 E+03 � m2

bbl × 1.589 873 E–01 � m3

ft × 3.048* E–01 � mft3 × 2.831 685 E–02 � m3

°F (°F–32)/1.8 � °Cpsi × 6.894 757 E+00 � kPa

*Conversion factor is exact.

Nikola Maricic is a petroleum engineer with the Chevron En-ergy Technology Company’s (ETC) reservoir simulation consult-ing team in Houston. E-mail: NMaricic@chevrontexaco.com.His research includes reservoir characterization and CBM simu-lation studies. Maricic has been working in the area of gas-storage reservoirs, heavy-oil production and operations, and,recently, black-oil reservoir simulation. He holds a BS degreefrom University of Belgrade (2002) in Serbia and an MS degreefrom West Virginia University (2004), both in petroleum engi-neering. Shahab D. Mohaghegh is a professor of petroleumand natural gas engineering at West Virginia University in Mor-gantown, West Virginia. E-mail: Shahab@wvu.edu. He has 17years of experience applying Artificial Intelligence and DataMining (AI&DM) to the upstream oil and gas industry includingall aspects of CBM reservoir engineering and production. Heholds BS and MS degrees in natural gas engineering from TexasA&M University and a PhD degree in petroleum and naturalgas engineering from Penn State University. Emre Artun is worksfor Chevron ETC’s reservoir simulation consulting team in Hous-ton. E-mail: EmreArtun@chevron.com. He holds a PhD degreefrom Penn State University, an MSc degree from West VirginiaUniversity, and a BSc degree from Middle East Technical Uni-versity, Turkey, all in petroleum and natural gas engineering.He also holds a graduate minor in computational science fromPenn State.

Fig. 22—NPVs for five isotherms after 15 years of production.Fig. 23—Gas-recovery results for different well configurationsand for different cases of thickness.

Fig. 24—NPV results for the case with three layers, each with athickness of 4 ft, for different well configurations.

983December 2008 SPE Reservoir Evaluation & Engineering