The discharge of mine water from underground coal mines in arid areas is leading to extensive water loss and secondary pollution. To eliminate the water loss and environmental pollution, sealing the concentrated brine after the mine water treatment in the underground storage and pumping the clean water to the surface for recycling are effective methods. In this study, focusing on the Ling Xin mining area, China, a coupled model of concentrated brine flow and solute transport in underground reservoir was established. The migration patterns of concentrated brine under two simulation scenarios of long-term penetration and sudden leakage were analyzed. At the same time, it also initially revealed the essence of the environmental pollution caused by the penetration and leakage of the concentrated brine in the underground concentrated brine storage reservoir. Results show that the concentrated brine would penetrate the bottom aquifer in about 60 years in long-term penetration while approximately 40 days in sudden leakage. In addition, the storage time, reservoir permeability, and groundwater head difference were important factors affecting the migration of concentrated brine, where the influence of permeability varieties was the most significant. The results of this study provide technical options for the subsequent study of the environmental risk of underground concentrated brine reservoirs and have important technical significance for the study and engineering application of underground reservoirs in arid areas.

Flow characteristics and phase distribution of concentrated brine storage in mining area have been core factors linking to the leakage risk assessment and ecological evaluation. Notably, saturation plays a crucial role in impacting the flow characteristics and distribution of brine, while the existence of oil left by mining machines and original reservoir and gas produced from coal bed gas and air has complicated the issue. In this work, we conducted the microfluidic visualization experiments to reveal the saturation distribution during brine storage in mining area. We applied machine learning model to extract saturation data from experimental images with over 95% accuracy. Eventually, we found that the existence of gas significantly impacts on the saturation distribution in micropores accounting for more than 80% contribution. We clarified that the gas production rate of median 200 μL/min impacts the least on saturation variation. Results in this research are of significance for deeper comprehension on three-phase saturation characteristics of concentrated brine storage in mining area.