Reservoir characterization of coals in the Powder River Basin, Wyoming, USA, for CO 2 sequestration feasibility studies

Transcription

1 Reservoir characterization of coals in the Powder River Basin, Wyoming, USA, for CO 2 sequestration feasibility studies Hannah E. Ross 1, Mark D. Zoback 1 1 Department of Geophysics, Stanford University, Stanford, CA , USA Abstract Coalbeds are an attractive geological environment for CO 2 sequestration because CO 2 is retained in the coal as an adsorbed phase and the cost of sequestration can be offset by enhanced coalbed methane recovery (ECBM). In order to examine the feasibility of sequestering CO 2 in unmineable coalbeds of the Powder River Basin (PRB), Wyoming, USA, we have carried out a reservoir characterization study and preliminary fluid flow simulations. We have focused our study on the Big George coal, part of the Wyodak-Anderson coal zone. Our 3D model of the Big George coal was built in an area of the PRB where the least principal stress is equal to the overburden stress, which would result in horizontal fracture propagation if hydraulic fracturing was utilized to enhance injectivity and recovery. We populated our 3D grid with cleat and matrix permeability and porosity data from the literature using geostatistical techniques and constrained the cleat permeability and porosity values through historymatching production data from CBM wells used to build our model. Simulations were run with pure CO 2 injection and with and without coal matrix shrinkage and swelling. We found that gravity and buoyancy are the major driving forces behind gas flow within the coal, and that coal matrix swelling results in a slight reduction in injectivity. However, hydraulically fracturing the coal close to its base helps mitigate the negative effect of permeability reduction on injection rate. Our simulations show that ~95% of the total CO 2 injected into the Big George coal would be sequestered and that CH 4 production would be ~6 times greater with CO 2 injection than without. We found that one injection well was able to sequester ~22 kt of CO 2 a year. Based on this injection rate, it would take ~2,3 injection wells (each with a lifetime of ~6 years) to sequester the current CO 2 emissions for the State of Wyoming. Since there have already been ~15, CBM wells drilled in the PRB, and ~5, more projected to be drilled in the next decade, utilization of 2,3 wells for CO 2 sequestration is quite feasible, especially in light of the potential for significant cost recovery through enhanced methane production. Keywords: CO 2 sequestration, unmineable coalbeds, enhanced coalbed methane, hydraulic fractures Introduction We examined the feasibility of sequestering CO 2 in unmineable coalbeds by conducting a reservoir characterization study and fluid flow simulations on coalbeds in the Powder River Basin (PRB), Wyoming, USA. In particular, we were interested in modeling the effects of horizontal hydraulic fractures on CO 2 injectivity and enhanced coalbed methane recovery (ECBM). Our study focused on the sub-bituminous Big George coal, part of the Wyodak-Anderson coal zone of the Tertiary Fort Union Formation. A 3D model of the Big George coal was constructed in an area of the PRB where the least principal stress is vertical, thereby guaranteeing horizontal hydraulic fractures. We built our model using well logs from coalbed methane (CBM) wells, and populated the model with permeability and porosity values using geostatistical techniques and history-matching. The PRB was chosen for several reasons: The basin is the location of the fastest growing natural gas play in the USA, mostly from the development of CBM from coalbeds in the Fort Union Formation (Figure 1). The USGS PRB 1

2 19W 47N 17W 16W 15W N 14W To ng ue 45N Montana Wyoming Sheridan Big Buffalo Ho Wyoming Study Area Casper Douglas 17W 5 16W 43N Coal Bed Methane Development Area Existing CO 2 pipelines km 18W lls nt Pla 44N Hi ns ute Sh To k ee Cr 45N k ac Bl tai un Mo 43N 46N Gillette rn 44N 42N 19W 13W 47N North Dakota Ri ve r 46N 18W South Dakota Po wd er Ri ve r Province Assessment Team [1] has estimated the total CBM resource in the PRB to be 14.3 trillion cubic feet (TCF). The state of Wyoming contains point sources for the capture of CO2, including several coalfired power plants in the south-western, eastern and north-eastern parts of the state that emit 1.7 TCF of CO2 per year [2]. The state of Wyoming has a CO2 pipeline network, with a proposed extension to the edge of the PRB (Figure 1) [2]. We collaborate with several research groups in the Geophysics and Petroleum Engineering departments at Stanford University, USA, who have ongoing projects in the PRB. The projects involve laboratory studies on coal samples from the PRB, monitoring CO2 migration in PRB coals, and categorizing fracture growth from water enhancement practices in PRB coals by CBM operators. The PRB meets most of the criteria set out in Gale and Freund [3] for identifying potential CO2 sequestration sites: Many of the coalbeds are laterally extensive and isolated from surrounding strata by shale beds; the basin is not strongly faulted or folded; permeability in the coals is high enough for CBM production; and the coals fall within the optimal depth range (~35-15 m). km 1 15W Possible future extension W 42N 13W Figure 1 Location map of the Powder River Basin, Wyoming, and our study area [modified from 4]. Our 3D model is located in the southern part of our study area. Reservoir characterization We have focused our study on the Big George coal, which is located in the central part of the PRB and is an amalgamation of five coalbeds [5]. The average depth of the Big George coal is 335 m and it varies in thickness from 14 to 62 m. Water enhancement tests and gamma ray logs have been used to characterize the Big George coal in our study area. Gamma ray logs from CBM wells gave us the depth and thickness of the coal, and water enhancement tests were analyzed to determine the direction of hydraulic fracture propagation in the coal. Water enhancement tests are used by CBM operators in the PRB to connect the CBM wells to the natural coalbed fracture network. In some areas the hydraulic fractures are found to propagate vertically, whereas in others they grow horizontally [4]. Knowing where vertical hydraulic fractures will form in the coal is especially important when choosing a site for 2

3 CO 2 sequestration because hydraulic fractures that propagate vertically may penetrate the overlying strata creating potential leakage conduits for CO 2. We constructed a 3D model of the Big George coal in an area of the PRB where the least principal stress (S 3 ) is the overburden stress, which means that hydraulic fractures will propagate horizontally in this area [4]. In our model the Big George coal is approximately 16 m thick and ranges in depth (to the top) from m, with a slight dip to the west (Figure 2). The number of grid blocks in our model is 1332 (42 x 41 x 6). We used sequential Gaussian simulation and simple kriging to populate our 3D model with multiple cleat and matrix permeability and porosity realizations (to capture heterogeneity and anisotropy) [6]. Our initial permeability and porosity values came from the literature [7, 8, 9, 1, 11] and we have further constrained the cleat permeability and porosity values through historymatching production data from the active CBM wells used to build our 3D model (Figure 2) mD mD m m 1 a) b) c) N mD Figure 2 a) Horizontal face cleat permeability. b) Horizontal butt cleat permeability. c) Vertical face cleat permeability. This figure shows our 3D model populated with cleat permeability values for realization 1. The heterogeneity and anisotropy in coal cleat permeability is modeled using geostatistical techniques. The horizontal face cleat permeability is higher than in the butt cleat or vertical directions [12]. Reservoir simulations Reservoir fluid flow simulations were run on a 5-spot well pattern (four producers and one injection well) with 8-acre well spacing using the Computer Modelling Group s ECBM simulator GEM. Our model and simulation input parameters are listed in Table 1. At present we have run all our simulations with pure CO 2 gas injection, with and without coal matrix shrinkage and swelling, and with and without a horizontal hydraulic fracture placed at the base of the injection well. Coal matrix shrinkage and swelling is modeled by the Palmer and Mansoori [13, 14, GEM 25] equation included in GEM, and the horizontal hydraulic fracture is modeled as a square fracture with dimensions 6 m x 6 m, porosity of 3%, and permeability of 1 md. In addition, we used gas adsorption isotherms specific to PRB coal. Our base case simulation is of CBM production for 11 years with no CO 2 injection, whereas subsequent simulations include CO 2 injection after 5 years of CBM production, with a total simulation time of 11 years. To prevent accidental hydraulic fracturing of the coal near the injection well, we set the maximum value of the bottom hole pressure (BHP) to less than 62 kpa, the fracture pressure in this area [4]. 3

4 Table 1 Input parameters for the fluid flow simulations. Input Parameters Values References Reservoir pressure gradient, 7.13 [15] kpa/m Injector and producer BHP 4, 1* *History-matching constraints, kpa Cleat spacing, cm 1 [8, 11] Matrix permeability, md.4-.7 [11] Matrix porosity [15] Cleat permeability, md Horizontal face cleat direction, 7-152, horizontal butt cleat direction, 1-48 and vertical direction, 1-48 Literature [7, 8, 9, 1, 11, 12] and history-matching Cleat porosity Literature [7, 9, 1, 15] and historymatching Adsorption/desorption parameters Langmuir volume:.577 gmole/kg for [16] for PRB coal samples (dry coal desorption for CH 4 and N 2, and moist coal adsorption for CO 2 ) CH 4, 1.67 gmole/kg for CO 2 Inverse Langmuir pressure, 1.7E-3/kPa for CH 4, 8.5E-4/kPa for CO 2 Diffusion coefficient, cm 2 /s.1 (1 days) for CH 4 and CO 2 [17] Rock compressibility Rock compressibility, 1.45E-7/kPa for matrix and 2.9E-5/kPa for cleats [18] for matrix and [7] for cleats Shrinkage/swelling [13, 14, GEM 25] Langmuir pressure for CH 4, 269 kpa, CO 2, 345 kpa [13, 14, 19, 2] Young s modulus,.413e7 kpa Poisson s ratio,.39 Matrix volume strain for CH 4,.7, CO 2,.13 Exponent, 3 Results and discussion Our fluid flow simulations of CO 2 injection into the Big George coal of the PRB show that gravity and buoyancy are the major driving forces behind gas flow within the coal. Gravity and buoyancy caused the gas to migrate upwards at first and then along the top of the coal, which reduced gas sweep efficiency and sequestration. We also find that coal matrix swelling results in a very slight reduction in CO 2 injectivity, but that hydraulically fracturing the coal close to its base mitigates the negative effect of permeability reduction on injection rate (Figure 3). Placement of a hydraulic fracture at the base of the injection well increased the total volume of CO 2 injected into the coal by ~3% (Figure 3). In regards to ECBM, CH 4 production was ~6 times greater with CO 2 injection than without (Figure 3). In order to model the expected decrease in CO 2 injectivity due to matrix swelling we used a multiphase composite version of the Palmer and Mansoori equation [13, 14, GEM 25]. However, matrix swelling had very little effect on CO 2 injectivity and CH 4 production, as shown in Figure 3. We find that the high cleat compressibility for the Big George coal dominates the modified Palmer and Mansoori equation so that the linear elastic term (which incorporates cleat compressibility) is comparable in magnitude to the matrix shrinkage and swelling term (which incorporates matrix volume strain due to the adsorption and desorption of gases). Decreasing the cleat compressibility would mean that the matrix shrinkage and swelling term had a more significant effect on injectivity. The matrix shrinkage and swelling term would also be more important in the permeability and porosity calculations if higher matrix volume strains were used. We are currently running laboratory experiments on coal 4

5 samples from the PRB to characterize coal matrix shrinkage and swelling. Total Volume of CH4 Produced (MSCF) Total Volume of CO2 Injected (tonne) CBM CBM only Tot al Volume of CO 2 Injected Injection, no hydrofrac, no Tot al Volume of CH 4 Produced ECBM Injection, no hydrofrac, no Injection, with Injection, with hydrofrac, with hydrofrac, with Figure 3 Total volume of CO 2 injected and total volume of CH 4 produced after 11 years. Hydraulically fracturing the coal at the base of the injection well increased the total volume of injected CO 2 by ~3%. With ECBM there was a ~6 fold increase in CH 4 production. Hydrofrac stands for hydraulic fracture and stands for matrix shrinkage and swelling. Conclusions To determine the feasibility of sequestering CO 2 in unmineable coalbeds of the PRB we have carried out a reservoir characterization study and fluid flow simulations. We used geophysical and geological data from the PRB to develop a 3D model of the Big George coal, and used geostatistical techniques to populate our model with numerous coal cleat and matrix permeability and porosity realizations. Initial results from fluid flow simulations show that gravity and buoyancy drive gas migration and matrix swelling slightly reduces gas injectivity. However, placing a horizontal hydraulic fracture at the base of the injection well helps to overcome the negative effect of matrix swelling on injection rates. Our simulations suggest that we can sequester ~95% of the total CO 2 injected into the Big George coal (increasing the well spacing would increase this number), that CH 4 production will be ~6 times greater with CO 2 injection than without, and that one injection well will be able to sequester ~22 kt of CO 2 a year. Based on this injection rate, it would take ~2,3 injection wells (each with a lifetime of ~6 years) to sequester the current CO 2 emissions for the State of Wyoming. Acknowledgements We thank the Stanford Global Climate and Energy Project for funding this project. References - [1] U. S. Geological Survey Powder River Basin Province Assessment Team. Executive Summary: Assessment of Coalbed Gas Resources of the Powder River Basin Province, Wyoming and Montana. U. S. Geological Survey Digital Data Series 24; DDS-69-C. - [2] Nummedal D, Towler B, Mason C, Allen M. Enhanced oil recovery in Wyoming: Prospects 5

6 and challenges: prepared for Governor Dave Freudenthal. University of Wyoming 23; < - [3] Gale J, Freund P. Coal-bed methane enhancement with CO 2 sequestration worldwide potential. Environmental Geosciences 21; 8: [4] Colmenares LB, Zoback MD. Hydraulic fracturing and wellbore completion of coalbed methane (CBM) wells in the Powder River Basin, Wyoming: Implications for water and gas production. AAPG Bulletin in review. - [5] Flores RM, Bader LR. Fort Union coal in the Powder River Basin, Wyoming and Montana: A synthesis. U.S. Geological Survey Professional Paper 1999; 1625-A:PS8-PS29. - [6] Deutsch CV. Geostatistical reservoir modeling: Oxford University Press; [7] U. S. Geological Survey National Oil and Gas Resource Assessment Team. National assessment of United States oil and gas resources. U. S. Geological Survey Circular 1995; [8] Ayers W. Coalbed gas systems, resources, and production and a review of contrasting cases from the San Juan and Powder River basins. AAPG Bulletin 22; 86: [9] Mavor MJ, Russell B, Pratt TJ. Powder River Basin Ft. Union coal reservoir properties and production decline analysis. SPE [1] Twombly G, Stepanek SH, Moore TA. Coalbed methane potential in the Waikato Coalfield of New Zealand: A comparison with developed basins in the United States. 24 New Zealand Petroleum Conference Proceedings. - [11] Flores RM. Coalbed methane in the Powder River Basin, Wyoming and Montana: An assessment of the Tertiary-Upper Cretaceous coalbed methane total petroleum system. U.S. Geological Survey Digital Data Series 24; DDS-69-C: chapter 2. - [12] Laubrach SE, Marrett RA, Olson JE, Scott AR. Characteristics and origins of coal cleat: A Review. International Journal of Coal Geology 1998; 35: [13] Palmer I, Mansoori J. How Permeability Depends on Stress and Pore Pressure in coalbeds: A New Model. SPE [14] Palmer I, Mansoori J. How Permeability Depends on Stress and Pore Pressure in Coalbeds: A New Model. SPE [15] Advanced Resources International, Inc. Powder River Basin coalbed methane development and produced water management. DOE/NETL-23/ [16] Tang G-Q, Jessen K, Kovscek AR. Laboratory and simulation investigation of enhanced coalbed methane recovery by gas injection. SPE [17] Stanford University Global Climate and Energy Project. Technical Report 23-24; [18] Law DH S, van der Meer LGH, Gunter WD. Comparison simulators for greenhouse gas sequestration in coalbeds, part III: More complex problems. 2 nd Annual Conference on Carbon Sequestration, May [19] Jones AH, Bell GJ, Schraufngel RA. A review of the physical and mechanical properties of coal with implications for coalbed methane well completion and production: Coal-bed Methane, San Juan Basin. Rocky Mountain Association of Geologists 1998; [2] Harpalani S. Gas flow characterization of Illinois coal: Final technical report. ICCI project number 3-1/7.1B