CO2 storage and enhanced coalbed methane recovery: Reservoir characterization and fluid flow simulations of the Big George coal, Powder River Basin, Wyoming, USA
Introduction
The amount of carbon dioxide (CO2) in the atmosphere has risen from pre-industrial levels of 280 ppm to present levels of approximately 380 ppm (Tans, 2007). This increase in atmospheric CO2 is attributed to the world's expanding use of fossil fuels and is believed to be one of the primary causes of global warming (Mann et al., 1998, EIA, 2006, IPCC, 2007). As a means for reducing greenhouse gas emissions, it has been proposed that CO2 be stored in geological formations, including mature oil and gas fields, deep saline aquifers and unmineable coal seams (IPCC, 2005). Of the three geological storage options, coalbeds are an attractive geological environment for CO2 storage because CO2 is retained in the coal as an adsorbed phase and the cost of storage can be offset by enhanced coalbed methane recovery (ECBM) (IPCC, 2005).
Stevens and Spector (1998) carried out a global assessment of the CO2 storage potential of unmineable coalbeds for the International Energy Agency Greenhouse Gas R&D Programme. They estimated the world coalbed CO2 storage capacity to be 114 trillion cubic meters (m3) (225 gigatons (Gt)). Their estimate was based on the assumption that total CH4 in-place would be replaced by CO2 at a ratio of 2:1. Parson and Keith (1998) carried out a similar study and estimated that the total storage capacity of unmineable coalbeds was approximately 185–556 trillion m3 (approximately 366–1100 Gt CO2), whereas Gale (2003) estimated the capacity to be 20 trillion m3 (40 Gt CO2). Later research has shown that the exchange ratios used by these studies vary enormously depending on the coal ranking (Stanton et al., 2001, Gluskoter et al., 2002) and such capacity estimates should be considered a first approach in estimating the storage potential of a coal-bearing basin.
In addition to rank, coal is also highly variable in terms of physical (porosity, permeability) and chemical properties (adsorption, diffusion). This complexity poses challenges for reservoir simulation packages designed to model conventional natural gas reservoirs. Law et al., 2002, Law et al., 2003, Law et al., 2004 undertook a comparison study of numerical CBM simulators to test whether they incorporated the features required to model gas fluid flow and adsorption, and to identify areas needing improvement for modeling ECBM (Law et al., 2002). Eight numerical simulators were used in the study, including GEM, ECLIPSE, COMET, SIMED II, GCOMP, METSIM 2, MoReS and COALCOMP. There were four parts to the study, the first focused on a single well test and pure CO2 injection into a 5-spot pattern model. The second was an extension of the first, but with CO2-enriched flue gas injection. The third was more complex, testing diffusion between the coal matrix and cleats and the effect of matrix shrinkage and swelling on gas adsorption. The final part involved history-matching field-test data collected by the Alberta Research Council (ARC) pilot project during injection of pure CO2 and flue gas into coalbeds in Alberta, Canada. In general there was very good agreement between the results (production well bottom hole pressure (BHP), injection well BHP, gas production rates, gas production compositions and gas saturation maps) from the different simulators for all four parts of the comparison study (Law et al., 2002, Law et al., 2003, Law et al., 2004).
In 2003, Taillefert and Reeves (2003) developed a screening model to assist industry in CO2 storage project consideration and screening. This screening model is used to predict the performance of ECBM projects under a number of reservoir conditions and operating assumptions. The model assumes that there is an existing CBM field that is being converted for ECBM recovery and CO2 storage. The model consists of a database of nearly 2000 reservoir simulation cases, which users can retrieve and compare. Users can choose between three values for each of the seven input parameters, including permeability, coal rank, depth, well spacing, injection rate, injection gas and injection timing. Coal thickness can be specified; although this is treated as a scaling parameter that controls total volumes. Both the model geometry (flat, horizontal layers) and physical properties, including porosity and fracture spacing, are fixed. Reeves et al. (2004) integrated a more robust ECBM and storage economic prediction module into Taillefert and Reeves’ (2003) screening model. A number of sensitivity analyses were conducted by varying permeability (using constant values for the entire reservoir), injected gas composition, well spacing, coal depth, injection rates and injection timing. The major finding was that ECBM operations are more favorable in low permeability, high rank coals (assuming no matrix shrinkage and swelling), because less CH4 is recovered during the primary production stage in these types of coals, compared to lower rank coals, and therefore CO2 injection is required to sweep the remaining CH4.
In contrast, Jackson (2006) found that ECBM was more beneficial in under-saturated, low rank, high permeability coals than in fully saturated coals because the presence of CO2 caused CH4 that normally would not desorb during de-pressurization in under-saturated coals, to desorb and be produced. Jackson argues that low rank, high permeability coals are good candidates for storage because the initial high permeability means that permeability reduction from matrix swelling will be less severe than in low permeability coals and therefore continued stimulation techniques will not be required for injection, reducing cost and time.
This paper presents a feasibility study to determine the CO2 storage potential of coalbeds in the Powder River Basin (PRB), Wyoming, USA, through geomechanically constrained, basin-specific fluid flow simulations (Fig. 1). The Powder River 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 (Fig. 1) (U.S. Department of Energy (DOE)/National Energy Technology Laboratory (NETL), 2003). The U.S. Geological Survey (USGS) Powder River Basin Province Assessment Team (2004) has estimated the total CBM resource in the PRB to be 416 billion m3 (14.3 trillion cubic feet (ft3)).
The coalbeds of the PRB satisfy many of the technical criteria outlined by Gale and Freund (2001), Bachu (2007) and Bachu et al. (2007) for identifying suitable coalbeds for CO2 storage. Coalbeds in the PRB have relatively high cleat permeabilities (> 1–5 mD), are at depths between 300 and 1000 m, are confined locally, and do not appear to have been deformed through faulting (Advanced Resources International, 2002). The high coal cleat permeability suggests that CO2 injection rates should be close to the CO2 delivery rates from large stationary point sources (Bachu, 2007). In addition, the development of both oil and CBM in the PRB means that infrastructure that has been put in place by the energy industry could be utilized for CO2 storage projects. At present the state of Wyoming has a CO2 pipeline network used for the transport of natural CO2 for enhanced oil recovery, with a proposed extension to the edge of the PRB (Fig. 1) (Nummedal et al., 2003). In the future, this pipeline could be converted for anthropogenic CO2 transportation to the basin. The state of Wyoming also contains various large point sources for the capture of CO2, including several coal-fired power plants in the southwestern, eastern and northeastern parts of the state that emit 23.8 billion m3 (47 million t) of CO2 per year (EIA, 2007).
The study presented in this paper focuses on the sub-bituminous Big George coal, part of the Wyodak-Anderson coal zone of the Tertiary Fort Union Formation (Fig. 2). A 3D stochastic reservoir model of the Big George coal was constructed and populated with permeability and porosity values using geostatistical techniques and history-matching. The main focus of this study was on gas buoyancy and leakage associated with CO2 injection in coalbeds. Because the Big George coal has a relatively high cleat permeability compared to the rate of diffusion, leakage due to buoyancy effects becomes an issue. To investigate the possibility of gas migration into overlying strata, several leakage scenarios were implemented in the reservoir simulator. However, for completeness, results are reported on the potential volumes of CO2 that could be stored and CH4 that could be produced from the Big George coal and on assessing the possible effects of matrix shrinkage and swelling and hydraulic fractures on injection rates. The effects of heterogeneity and vertical grid resolution on volume estimates were also investigated by constructing models with a homogenous cleat permeability field and only one model layer in the vertical direction.
Section snippets
Geologic setting and coal characterization
The Powder River Basin is located in southeast Montana and northeast Wyoming, and is an asymmetrical syncline enclosed by the Bighorn Mountains in the west, the Miles City Arch in the northeast, the Black Hills in the east, the Hartville Uplift in the southeast, and the Casper-Arch-Laramie Range in the southwest (Fig. 1). The axis of the basin is close to its western side (NW–SE), with the eastern flank dipping gently to the west at 2–5° and the western flank dipping to the east at 20–25° (
Lab and field data sources specific to the Powder River Basin
Field data collected from coalbeds in the PRB are extremely limited. For this study the authors had access to gamma ray logs from several active CBM wells (WOGCC), adsorption isotherms measured on PRB coal samples (Tang et al., 2005, Stricker et al., 2006) and stress measurements from hydraulic fracturing data obtained from CBM operators in the PRB (Colmenares and Zoback, 2007).
The stress state of coalbeds in the PRB was determined by Colmenares and Zoback (2007) using hydraulic fracturing data
Simulation scenarios
The reservoir fluid flow simulations were run with a 5-spot well pattern (four production wells at each corner and one injection well in the center) on a 0.32 km2 (80 acre) well spacing using the dual-permeability, single-porosity Coalbed Methane package of the Computer Modeling Group's Generalized Equation-of-State Model Compositional Reservoir Simulator (GEM, 2005). The 3D simulation model of the Big George coal used in this study is approximately 16 m thick and ranges in depth (to the top)
Reference case: CO2 storage and ECBM production
The predicted cumulative volumes of CO2 that could be stored and ECBM that could be produced for the reference case, for a 0.6 km2 (160 acre) area of the Big George coal, are shown in Fig. 4. This figure shows the results from populating the simulation model with the different property distribution realizations described in Section 3.3 and then running the reference case on each of these differently populated models (to populate the model the first realization for each coal property (cleat and
Discussion
These simulation cases actually show the lower bound on gas migration and saturation in overlying units, as the Big George coal and other coalbeds in the PRB can have permeabilities close to 1000 mD in other parts of the basin (Flores, 2004, Mavor et al., 2003), which would mean that gas migration into overlying units could be even faster in those areas than the simulations show. While a much finer simulation grid would reveal a clearer picture of gas migration through the Big George coal and
Conclusions
This paper reports on a reservoir characterization study and fluid flow simulations that were carried out to determine the feasibility of storing CO2 in the Big George coal of the PRB. Geological data from the PRB were used to develop a 3D model of the Big George coal, and geostatistical techniques and history-matching were undertaken to populate the model with numerous coal cleat and matrix permeability and porosity realizations. Results from fluid flow simulations show that gravity and
Acknowledgements
The authors thank the Stanford University Global Climate and Energy Project for funding this research. The authors are also grateful for comments and suggestions by Dr. Stefan Bachu and two anonymous reviewers that greatly improved this manuscript.
References (65)
Carbon dioxide storage capacity in uneconomic coalbeds in Alberta, Canada: Methodology, potential and site identification
International Journal of Greenhouse Gas Control
(2007)- et al.
CO2 storage capacity estimation: methodology and gaps
International Journal of Greenhouse Gas Control
(2007) - et al.
Variation in permeability with lithotype and maceral composition of Cretaceous coals of the Canadian Cordillera
International Journal of Coal Geology
(1997) - et al.
Characteristics and origins of coal cleat: a review
International Journal of Coal Geology
(1998) - et al.
A bidisperse pore diffusion model for methane displacement desorption in coal by CO2 injection
Fuel
(2003) - et al.
The change in effective stress associated with shrinkage from gas desorption in coal
International Journal of Coal Geology
(2001) - et al.
Enhanced coalbed methane and CO2 storage in anthracitic coals—micro-pilot test at South Qinshui, Shanxi, China
International Journal of Greenhouse Gas Control
(2007) - Advanced Resources International, Inc., 2002. Powder River Basin coalbed methane development and produced water...
- Applied Hydrology Associates, Inc. and Greystone Environmental Consultants, Inc., 2002. Groundwater modeling of impacts...
Coalbed gas systems, resources, and production and a review of contrasting cases from the San Juan and Powder River basins
AAPG Bulletin
(2002)
Importance of fabric and composition on the stress sensitivity of permeability in some coals, Northern Sydney Basin. Australia: relevance to coalbed methane exploitation
AAPG Bulletin
Hydraulic fracturing and wellbore completion of coal bed methane (CBM) wells in the Powder River Basin. Wyoming: implications for water and gas production
AAPG Bulletin
Geostatistical Reservoir Modeling
Coal-bed methane enhancement with CO2 sequestration worldwide potential
Environmental Geosciences
Going underground: CO2 storage and emissions reduction
Power Engineer
Measurement of “rock properties” in coal for coalbed methane production
GEM 2005. 15 User's Guide
Adsorption of carbon dioxide and methane in low-rank coals and the potential for sequestration of carbon dioxide
Environmental Geosciences
Mechanics of hydraulic fracturing
Petroleum Transactions of the American Institute of Mining Metallurgical and Petroleum Engineers
CO2 enhanced methane and reduced water production in high permeability, under-saturated coal seams—a modelling study
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
Numerical simulator comparison study for enhanced coalbed methane recovery processes. Part I. Pure carbon dioxide injection
Cited by (65)
Upscaling poromechanical models of coalbed methane reservoir incorporating the interplay between non-linear cleat deformation and solvation forces
2023, International Journal of Solids and StructuresApparent diffusion coefficient for adsorption-controlled gas transport in nanoporous media
2022, Chemical Engineering JournalCoal permeability evolution triggered by variable injection parameters during gas mixture enhanced methane recovery
2022, EnergyCitation Excerpt :Huang et al. [37] investigated the optimal scheduling for CO2 enhanced coal bed methane production. Field experiments of CO2-ECBM have been done in Poland [38,39], Wyoming, USA [40], Netherlands [41], India [42] and South Qinshui, Shanxi, China [43]. The common findings show that the injected CO2 maintains the reservoir pressure but decreases the partial pressure of CH4 [44,45].
A systematic capacity assessment and classification of geologic CO<inf>2</inf> storage systems in India
2021, International Journal of Greenhouse Gas ControlSingle-well production history matching and geostatistical modeling as proxy to multi-well reservoir simulation for evaluating dynamic reservoir properties of coal seams
2021, International Journal of Coal GeologyCitation Excerpt :Coal seam gas production involves primary depletion (Moore, 2012; Pashin, 2010) or enhanced gas production by injecting a secondary gas (such as CO2) to strip methane from the coal (Ross et al., 2009; Packham et al., 2012; Zhou et al., 2013; Le et al., 2020) with simultaneous sequestration.