A Case Study on the Numerical Simulation of Enhanced Coalbed Methane Recovery

Abstract

A hypothetical pilot-scale project was studied to investigate the capabilities of existing coalbed methane (CBM) models in simulating the multicomponent gas sorption and diffusion in coalbeds for CO2-sequestration enhanced coalbed methane (ECBM) production. Further, the effects of coal properties, e.g.cleat permeability, porosity, cleat spacing, sorption isotherms of CH4 and CO2, and initial reservoir conditions, on CO2 breakthrough and methane recovery were investigated as well. The calculated results suggested that the multicomponent sorption equilibria described using IAS/Langmuir isotherm and extended Langmuir isotherm are quite different. The gas diffusion rate can not be accurately modeled using Warren & Root equation based on dual-porosity model (pseudo-steady state approach), which is so far the most popular approach used in CBM models to describe gas diffusion process. This indicates that a rigorous multicomponent gas diffusion model should be incorporated in CBM models to improve numerical simulation and prediction of CBM/ECBM production. Introduction The numerical simulation model is a useful tool in the development of the CO2-sequestration for coalbed methane recovery technology. Currently many simulation models have been developed to simulate the ECBM recovery process. However, a complete description of this process has proven to be challenging, primarily because of the extreme complexity of the process. This paper presents a comparison of two CBMmodels: EclipseCBM model and PSU-Coalcomp model. Both the models are three-dimensional, two-phase, and are capable to model the CBM and ECBM production, but neglecting the variation of fractureporosity and permeability due to coal shrinkage and swelling. The binary gas diffusion dynamics is modeled using a pseudo-steady state approach. The method treats matrix sorption as a pseudosteady-state transport process, which is similar to the formulation of Warren and Root for conventional naturally fractured reservoirs[1, 2]. For single-component gas transport, the equation is based on the Fick's first law of diffusion (1) where C is average gas concentration in bulk coal and Cpg is surface gas concentration. The sorption time constant t is a lumped parameter incorporating both diffusion time and rate of sorption. It regulates the rate at which gas is released from the micropores to natural fracture system by the following equation: (2) where Dc represents the micropore diffusion coefficient, s is geometric factor. The difference between these two models is the sorption isotherms used to describe sorption eqilubrium: the extended Langmuir isotherm used in Eclipse CBM model, and the IAS/Langmuir isotherm in PSU-Coalcomp. The case study was performedin two steps. Firstly, some unknown parametersweredeterminedbyfitting the calculated results of Eclipse model with those of PSU model [3] to constrcut a "corrected" Eclipse model for further study. The coal properties and initial conditions werethen altered to investigate the sensitivity of CO2 breakthrough and methane recovery to these parameters, and to analyze the capabilities of two models in modeling binary gas diffusion process. Model comparison The case demonstrated in a paper of O. Odusote et al. [3] was used for model comparison. A square well pattern with one vertical injector in the center of four horizontal producers was used to represent this hypothetical pilot-scale project, in which primary methane production took place for one year before CO2 injection-ECBM recovery is initiated. The production and injection wells are to be shut-in at the time of CO2 breakthrough, at which the mole fraction of CO2 in the gas production stream is equal to 5%.