Gas Pressure Prediction Model for Carbon Dioxide Injection in a Large-Scale Coal Matrix Experiment

Abstract

Carbon dioxide injection, gas flow, and adsorption of CO2 in a coal matrix are modelled by a large-scale physical experiment. The process primarily comprises an increase in the gas pressure from an unsaturated to a saturated status during gas injection followed by a reduction in the gas pressure due to the adsorption of CO2 after the cessation of gas injection. We developed a straightforward and effective calibration-based gas pressure model for the spatio-temporal prediction of CO2 in a coal matrix. The Nash–Sutcliffe efficiency of the models for predicting gas pressure is between 0.931 and 0.998. The results indicate that the spatio-temporal distribution of gas pressure in the coal matrix mainly relates to the gas pressure injected, geological boundary conditions, initial geo-stress, and gas type. The gas concentration tends to increase in high Earth stress and boundary areas during gas injection, whereas the gas adsorption rate is high in areas of low Earth stress. The gas adsorption rate of the coal matrix decreases as the gas injection content and gas injection time increase; improvement of the gas injection pressure can boost the gas adsorption to some extent. This model can explain or predict the behaviour of gas in a coal matrix depending on the matrix density, and it has the potential to simulate gas behaviour in a large-scale undisturbed coal sample considering the structure (joints and cleats) of the coal.