Impact of transition from local swelling to macro swelling on the evolution of coal permeability

Laboratory observations have shown that coal permeability under the influence of gas adsorption can change instantaneously from reduction to enhancement. It is commonly believed that this instantaneous switching of permeability is due to the fact that the matrix swelling ultimately ceases at higher pressures and the influence of effective stresses take over. In this study, our previously-developed poroelastic model is used to uncover the true reason why coal permeability switches from reduction to enhancement. This goal is achieved through explicit simulations of the dynamic interactions between coal matrix swelling/shrinking and fracture aperture alteration, and translations of these interactions to perrmeability evolution under unconstrained swellings. Our results of this study have revealed the transition of coal matrix swelling from local swelling to macro-swelling as a novel mechanism for this switching. Our specific findings include: (1) at the initial stage of CO ₂ injection, matrix swelling is localized within the vicinity of the fracture compartment. As the injection continues, the swelling zone is extending further into the matrix and becomes macro-swelling. Matrix properties control the swelling transition from local swelling to macro swelling; (2) matrix swelling processes control the evolution of coal permeability. When the swelling is localized, coal permeability is controlled by the internal fracture boundary condition and behaves volumetrically; when the swelling becomes macro-swelling, coal permeability is controlled by the external boundary condition and behaves non-volumetrically; and (3) matrix properties control the switch from local swelling to macro swelling and the associated switch in permeability behavior from reduction to recovery. Based on these findings, a permeability switching model has been proposed to represent the evolution of coal permeability under variable stress conditions. This model is verified against our experimental data. It is found that the model predictions are consistent with typical laboratory and in-situ observations available in lietratures.