A coupled thermal-nonlinear hydraulic-mechanical (THM) model for fault water inrush was carried out in this paper to study the water-rock-temperature interactions and predict the fault water inrush. First, the governing equations of the coupled THM model were established by coupling the particle transport equation, nonlinear flow equation, mechanical equation, and the heat transfer equation. Second, by setting different boundary conditions, the mechanical model, nonlinear hydraulic-mechanical (HM) coupling model, and the thermal-nonlinear hydraulic-mechanical (THM) coupling model were established, respectively. Finally, a numerical simulation of these models was established by using COMSOL Multiphysics. Results indicate that the nonlinear water flow equation could describe the nonlinear water flow process in the fractured zone of the fault. The mining stress and the water velocity had a great influence on the temperature of the fault zone. The temperature change of the fault zone can reflect the change of the seepage field in the fault and confined aquifer. This coupled THM model can provide a numerical simulation method to describe the coupled process of complex geological systems, which can be used to predict the fault water inrush induced by coal mining activities.

A coupled thermal–hydraulic–chemical (THC) model was carried out in this paper to study the influence of rock heterogeneity and the coupling effect of temperature, groundwater, and hydrochemistry on rock damage. Firstly, the hydrochemical and hydraulic erosion equations were established. The equations of the coupled THC model were established by combining the hydrochemical and hydraulic erosion equations, the flow equations, and the heat transfer equations. Weibull distribution was adopted to govern the heterogeneity of initial rock porosity distribution. Secondly, the influence of the hydrochemistry, the temperature and the initial porosity heterogeneity on porosity and fluid velocity change was studied. Then the rock damage rule changed with time at different pH values and temperature was studied. Finally, an actual deep coal mine model was established to apply the THC model to predict water inrush. Results indicate that: (1) The average porosity and average fluid velocity approximately show linear growth and exponential growth with time, respectively, and their growth rates increase with decreasing pH value and increasing temperature in a certain acidity and temperature range. (2) The increase of initial porosity heterogeneity has little influence on porosity change, but it can increase the fluid velocity growth rate. The porosity heterogeneity and fluid velocity heterogeneity approximately show exponential growth with increasing time, and the rock heterogeneity growth contributes to form cracks. The increase of temperature and decrease of pH value have little influence on the porosity heterogeneity, but they can increase the growth rate of the fluid velocity heterogeneity. (3) The rock damage shows linear growth with time, and its growth rate increases with decreasing pH value and increasing temperature in a certain acidity range and temperature range. (4) The increase of rock heterogeneity can increase the possibility of water inrush.

Water inrush is a serious geological disaster in deep coal mining and rock engineering in China. The occurrence of water inrush is affected by the interaction of many controlling factors in underground reservoir areas. In order to improve the accuracy of water inrush risk evaluation, the changes in index weights due to the change in influencing factors and the uncertain factors should be both considered. In this study, an evaluation model that integrates variable weight and unascertained measure (VWUM) theories was established. First, a modified credible degree recognition was established to make the evaluation results represent both the evaluation grade and the difference in the degree to which the indexes belong to the same grade. Then, the VWUM model was established by coupling the variable weight, unascertained measure, modified credible degree recognition, and geographic information system (GIS). Last, the VWUM model is applied to a case study to evaluate water inrush risk. Results indicates that more detailed risk grades and their spatial distribution zones can be obtained by the VWUM model compared with those obtained by the water inrush coefficient. The VWUM model can provide a more accurate evaluation result.