Modeling and simulation

for CBM production 9

Kashy Aminian
West Virginia University, Morgantown, WV, United States

9.1 Predicting the behavior of CBM reservoirs

The gas production from CBM reservoirs is governed by complex interaction of
single-phase gas diffusion through coal matrix and two-phase gas-water flow through
the cleat system that are coupled by the desorption process. CBM reservoirs are
heterogeneous and anisotropic porous media with complex pore structure. Addition-
ally, permeability of a coalbed is simultaneously impacted by changes in effective
stress, gas slippage, and matrix shrinkage during gas production. As a result, analyzing
and predicting the production behavior of CBM reservoirs is complex and challeng-
ing, especially at the early stage (dewatering) of the recovery. CBM reservoirs exhibit
a very different gas production profile as compared to conventional gas reservoirs.
Several approaches, both analytical and numerical, have been presented to predict
the production performance for coalbed methane reservoirs [1]. Analytical and
semianalytical models have been proposed to predict CBM production. However,
most of these models are only for single-phase flow of gas. Analytical techniques
are useful for estimating well performance of the mature producing fields with large
amount of reservoir data available [2]. It is difficult to model all complexities involved
in CBM production process by analytical methods. The complex storage and transport
mechanisms involved in CBM production can be best represented by the numeri-
cal models. Numerical models are commonly employed for evaluation of CBM
production [3].

9.2 Reservoir simulation

Reservoir simulation is a useful technology for solving complex reservoir engineering
problems with efficiency. The numerical models are considered the best tool for CBM
reservoir development. Numerical models can accurately and simultaneously account
for gas desorption, relative permeability, permeability, and porosity change caused
by the coal compressibility, permeability changes due to matrix shrinkage, well-to-
well interference, and operating procedures. Reservoir simulators can be used to
perform a variety of analyses. The primary uses of simulators include predicting
production performance of the CBM reservoirs under various reservoir management

Coal Bed Methane. https://doi.org/10.1016/B978-0-12-815997-2.00009-3

© 2020 Elsevier Inc. All rights reserved.
170 Coal Bed Methane

strategies, estimating the ultimate gas recovery, and designing the most effective well
completions. As result, simulation has become an indispensable tool for reservoir
management because of the need for support decisions and economic justification
with accurate technical data. Detailed information for perform a CBM simulation
study and guidelines for history matching and production forecasting are documented
by Paul [4].
The complex nature of gas production from CBM reservoirs requires significantly
more reservoir parameters than conventional reservoirs. Often, all the required
reservoir parameters needed to evaluate a CBM prospect are not available. Further-
more, many samples are often required to establish representative values for various
parameters because of the heterogeneous nature of the coal. In absence of sufficient
data, parametric studies are often conducted to evaluate the impact of the missing
parameter on the recovery, well performance, and future revenues. Reservoir simula-
tion is a cost-effective way to assess the sensitivity of the reservoir performance to
uncertainties in measured data.

9.3 CBM reservoir simulators

A number of CBM reservoir simulators have been described in literature. King and
Ertekin [3] provided an excellent review of the development of numerical simulator
for CBM before the late 1980s. Wei et al. [5] provided a review of recent advances
in numerical simulation for primary CBM recovery and enhanced CBM (ECBM)
recovery. The simulators can be classified into two general categories by the way they
treat the gas sorption process, i.e., equilibrium and nonequilibrium. The equilibrium
models are conventional single-porosity models that have been converted to model
CBM reservoirs. In these types of models, two-phase fluid flow through fracture net-
work in CBM reservoirs is modeled according to Darcy’s law. A pressure-dependent
source term is used in these models to account for gas desorption and the gas diffusion
is not considered [6].
The equilibrium models ignore the sorption time, which usually is not a problem,
because for most coals the gas desorption rate is rapid enough such that gas flow rates
are limited fracture system permeability [7]. However, for coals with longer desorp-
tion times, the nonequilibrium simulators may not result in realistic predictions of the
gas recovery. These models are dual-porosity simulators modified for gas storage by
adsorption and gas transport by diffusion in the matrix. The diffusion coefficient can
be assumed to be independent of gas concentration (pseudo-steady-state formulation)
or concentration dependent (unsteady-state formulation). The unsteady-state simula-
tors are the most rigorous, because they include the effects of the gas concentration
gradient. However, they also are computationally taxing because of the complexity
of the solution. Comparison of the two approaches shows that although the pseudo-
steady-state method underestimates gas rates in the first 100 h or so, results at longer
times are virtually identical [7].
Modeling and simulation for CBM production 171

To model changes in produced gas composition with time [8] and injection of
miscible gases such as nitrogen or carbon dioxide for ECBM recovery [9], composi-
tional simulators must be considered. In this approach, additional capabilities to
account for the multicomponent adsorption-desorption kinetics and mixing of the
injected gas and desorbed methane are required. However, because of extreme com-
plication of physical and thermodynamic phenomena that occur during CBM/ECBM
recovery process, the existing models do not adequately model the mechanism of
gas recovery from coal.
Recent studies suggest that lower rank coals exhibit a multimodal pore structure
[10, 11]. Correspondingly, unipore and bidisperse diffusion models are developed
to model different types of gas diffusion processes. The bidisperse models generally
assume a two-step gas diffusion in coal matrix. Gas adsorption takes place in the
micropores with the mesopores or macropores providing storage for free gas, as
well as tortuous paths for transport of gas between the micropores and cleats. In these
bidisperse models, the gas diffusion in micro- and mesopores is assumed to be
controlled by different mechanisms. Many attempts have been recently made to
improve the CBM/ECBM recovery simulation with alternative models such as the
bidisperse pore diffusion model and the triple porosity model [11–14]. In these
models, the gas diffusion in the coal matrix was described as bidisperse diffusion.
Additionally, several empirical equations were incorporated in the simulator to
describe variation of coal properties because of coal shrinkage. However, a number
of issues remain unresolved in rigorously model CBM production, including represen-
tation of multiscale pore structure, multicomponent gas diffusion, and impact of
water/moisture on gas diffusion in coal.

9.4 Sensitivity analyses

One of the main applications of the reservoir simulation is to conduct sensitivity or
parametric studies to assess the impact of uncertain or missing data on CBM reservoir
performance and to evaluate the economics of various operating procedures. Paramet-
ric simulations can be performed by varying certain data over their expected ranges to
gain a general sense of the potential for gas recovery. Such an approach might be used
before drilling in a step-out or previously undeveloped area to assess the commercial
viability of producing coalbed methane reserves. Gas recovery forecasts will be highly
speculative in such cases. When some measured data are available but show consid-
erable scatter, parametric simulations over the range of measured values can help to
determine the tolerance for error in the measurements and which data are the most
critical to assessing reservoir performance [15]. Usually, these data will include cleat
permeability, gas content, and the adsorption isotherm. A number of parameters can
be varied in the simulations to optimize field operating procedures [16]. These param-
eters might include pressure and rate constraints, well backpressure or bottomhole
pressure schedules, well location and spacing, well completion practices for new
172 Coal Bed Methane

wells, and recompletion practices for existing wells. This type of sensitivity analysis
usually is performed once a history match has been completed. This analysis is one of
the more powerful uses of reservoir simulation because it enables forecasts beyond
history to be generated.
A pilot test is often designed and implemented to verify the productivity poten-
tial of the CBM reservoirs. The objective of the pilot test program is to determine
gas deliverability. Estimating water production is also important for water disposal
considerations. A secondary purpose is to confirm the gas content and permeability
estimates from initial data wells. The pilot program production and test data should
be used to estimate hard-to-define reservoir parameters, including permeability
anisotropy, relative permeability characteristics, fracture system porosity, pore
compressibility, and initial water saturation by performing simulation study and
history matching. These parameters can be then used to accurately predict long-
term production performance. History matching the pilot pattern data should also
resolve uncertainties in key reservoir data, such as desorption isotherms and gas
content data that may arise due to experimental error and normal data scatter
[14]. Parametric simulations can then be performed to evaluate the effect of
variations in reservoir properties on production. Figs. 9.1–9.3 show the effects
of spacing, permeability, and hydraulic-fracture length, respectively. Anything that
the operator can do to promote greater interference with offset wells will promote
greater desorption and lead to quicker and higher gas production. The economics
will dictate what the optimum spacing and pattern should be, given the different
scenarios.

Fig. 9.1 Production rates for various permeabilities.

Modeling and simulation for CBM production 173

Fig. 9.2 Production rates for various fracture half-lengths.

Fig. 9.3 Production rates for various well spacing.

References
[1] Gerami S, Pooladi-darvish M, Morad K, Mattar L. Type curves for dry CBM reservoirs
with equilibrium desorption. J Can Pet Technol 2008;47(7):49–56.
[2] Zuber MD. Basic reservoir engineering for coal. In: A guide to coalbed methane reservoir
engineering. Chicago, Illinois: Gas Research Institute Report GRI-94/0397; 1996.
174 Coal Bed Methane

[3] King GR, Ertekin T. A survey of mathematical models related to methane production from
coal seams. In: Paper presented at the coalbed methane symposium, Tuscaloosa, Alabama,
April 17–20; 1989.
[4] Paul GW. Simulating coalbed methane reservoirs. In: A guide to coalbed methane reser-
voir engineering. Chicago, Illinois: Gas Research Institute Report GRI-94/0397; 1996.
[5] Wei XR, Wang GX, Massarotto P, Golding SD, Rudolph V. A review on recent advances
in the numerical simulation for coalbed-methane-recovery process. SPE Reserv Eval Eng
2007;10:657–66.
[6] Manik J. Compositional modeling of enhanced coalbed methane recovery. PhD disserta-
tion, Penn State University; 1999.
[7] Kolesar JE, Ertekin TM, Obut ST. The unsteady-state nature of sorption and diffusion
phenomena in the micropore structure of coal: Part 2—Solution. SPE Format Eval
1990;5:89–97.
[8] Deo MD, Whitney EM, Bodily DM. A multicomponent model for coalbed gas drainage.
In: Presented at the 1993 international coalbed methane symposium, The University of
Alabama/Tuscaloosa; May 17–21, vol. 1; 1993. p. 223–31.
[9] Arri LE, et al. Modeling coalbed methane production with binary gas sorption. In: SPE
paper 24363, Presented at the SPE Rocky Mountain regional meeting, Casper, Wyoming;
April 10–12; 1992.
[10] Gan H, Nandi SP, Walker PL. Nature of the porosity in American coals. Fuel 1972;51(4):
272–7.
[11] Clarkson CR, Bustin RM. The effect of pore structure and gas pressure upon the transport
properties of coal: a laboratory and modeling study. Fuel 1999;78(11):1333–44.
[12] Cui XJ, Bustin RM, Dipple G. Selective transport of CO2, CH4, and N2 in coals: insights
from modeling of experimental gas adsorption data. Fuel 2004;83(3):293–303.
[13] Mazumdur S, Plug WJ, Bruining J. Capillary pressure and wettability of coal-water-
carbon dioxide system. In: SPE paper 84339, Presented at SPE annual technical confer-
ence and exhibition, Denver, Colorado; October 5–8; 2003.
[14] Shi JQ, Durucan S. A bidisperse pore diffusion model for methane displacement desorp-
tion in coal by CO2 injection. Fuel 2003;82(10):1219–29.
[15] Zuber MD, Olszewski AJ. Coalbed methane production forecasting: measurement
accuracy required for key reservoir properties. In: Presented at the 1993 international
coalbed methane symposium, The University of Alabama/Tuscaloosa, May 17–21,
vol. II; 1993. p. 549–59.
[16] Saulsberry JL, Schraufnagel RA. Sensitivity of permeability and other reservoir properties
on coalbed methane gas recovery. In: Presented at the 1993 International Coalbed
Methane Symposium, The University of Alabama/Tuscaloosa, May 17–21, vol. I;
1993. p. 123–30.