Numerical investigation on the coupled gas-solid behavior of coal using an improved anisotropic permeability model

Coalbed methane (CBM) extraction uses coal seam depressurization to extract methane adsorbed into the coal matrix. The process involves complex interactions between solid and gas, and between mechanical and chemical effects. Permeability of coal is one of the most important parameters to evaluate the production of CBM reservoirs. Several models have been proposed to establish the relationship between permeability and poromechanical response, gas desorption, and coal shrinkage. However, there have been only a few attempts in describing the anisotropic permeability of coal, which is an important characteristic of gas flow in coal. This paper focuses on the anisotropic properties of coals and its effects on the coupled gas-solid behavior. We develop an improved anisotropic permeability based on linear elastic poromechanical theory and then implement this model into a finite element package to study the effects of the anisotropic permeability on the gas transport in coal. Cases with three boundary conditions including uniaxial strain, constant volume, and constant confining stress are analyzed in details. Numerical simulation results show that the proposed model can well represent the anisotropic permeability of coal. Permeability is closely related to the mechanical properties of coal. We find that modulus reduction ratio is a key parameter to control the anisotropic characteristics. Different modulus reduction ratios alter directional permeability significantly. Anisotropic permeability changes with time and pressure during gas desorption and gas flow. Further, the magnitude of directional permeability variations depends on both the modulus and the boundary conditions. Permeability evolution is controlled by the change in effective stress and matrix shrinkage due to desorption. The effects of anisotropic permeability on the coupled gas-solid behavior of coal during gas desorption and flow processes are demonstrated in this work. (C) 2016 Elsevier B.V. All rights reserved.