For underground water reservoirs in coal mines, the complex water-rich environment and changing overburden stress can damage coal pillar dams. In this paper, the coal samples from coal seam 2² of Shangwan coal mine were taken as research objects and the damage mechanism and characteristics of coal samples with different moisture content and wetting-drying cycles under cyclic loading were investigated. The results show that as the moisture content and wetting-drying cycles increase, the post-peak stage of the coal samples under cyclic stress becomes obvious, and the hysteresis loop changes from dense to sparse. Compared to the uniaxial compression experiment, when w = 5.28% (the critical water content), mechanical parameters such as peak strength and modulus of elasticity decrease the most. Under cyclic loading, the damage mode of both sets of coal samples was tensile damage, but the increase in wetting-drying cycles promotes the development of shear fractures. For evaluating fracture types, the RA-AF density map is more applicable to wetting-drying cycle coal samples, whereas for the coal samples with different moisture contents this should be carried out with caution. This study can provide some theoretical basis for the stability evaluation of coal pillar dams in underground water reservoirs.

Long-term erosion and repeated scouring of water significantly affect the technical properties of coals, which are the essential elements that must be considered in evaluating an underground reservoir coal column dam’s standing sustainability. In the paper, the coal pillar dam body of the 2² layers of coal in the Shangwan Coal Mine is studied (2² represents No. 2 coal seam), and the water content of this coal pillar dam body is simplified into two types of different water content and dry–wet cycle. Through acoustic emission detection technology and energy dissipation analysis method, the internal failure mechanism of coal water action is analyzed. This study revealed three findings. (1) The crest pressure, strain, and resilient modulus in the coal sample were inversely related to the water content along with the dry–wet cycle number, while the drying–wetting cycle process had a certain time effect on the failure to the sample. (2) As the moisture content and the dry–wet cycle times incremented, three features were shown: first, the breakage pattern is the mainly stretching fracture for the coal specimen; second, the number and absolute value of acoustic emission count peaks decrease; third, the RA-AF probability density plot (RA is the ratio of AE Risetime and Amplitude, and AF is the ratio of AE Count and Duration) corresponds more closely to the large-scale destruction characteristics for the coal samples. (3) A higher quantity of wet and dry cycles results in a smoother energy dissipation curve in the compacted and flexible phases of the crack, indicating that this energy is released earlier. The research results can be applied to the long-term sustainability assessment of the dams of coal columns for underground reservoirs and can also serve as valuable content to the excogitation of water-bearing coal column dams under similar engineering conditions.

Comprehending the water absorption process inherent to coal, including the associated spatial distribution patterns of water, proves indispensable in the design and evaluation of coal pillar dams in underground water reservoirs. To better understand this process, a series of NMR (nuclear magnetic resonance) tests were carried out on cylindrically shaped coal samples immersed in water for varying durations, with the upper and lower surfaces of the samples sealed. A method involving image digital processing and finite element simulation was used to quantitatively characterise the water absorption process, as well as the spatial distribution of water in the samples. The results showed that NMR imaging colour brightness differences were positively correlated with water content and that the wetted ring gradually increased in width as the water immersion time increased. The expectation and sum of squared deviations of the pixel greyscale values of the NMR images, which were used to characterise the water saturation and spatial distribution of the coal samples, represented positive and negative exponential functions of the water immersion time, respectively. This indicated that the water saturation gradually increased and became more uniformly distributed. Furthermore, based on the set threshold value of the target variable rate of change, the limiting expectation of the pixel greyscale values was obtained, and the limiting water absorption time of the coal sample was predicted. The water diffusion equation was then used to characterise the water absorption process of the coal samples, and a water diffusion model was developed to accurately obtain the wet ring boundary data. A reasonable value of the diffusion coefficient was determined by comparing and correcting the results of the numerical simulation and physical experiments with full consideration of the non-homogeneity of the numerical model. This water diffusion model can better characterise the water transport phenomena in the macroscopic barrier zone of coal pillar dams. Finally, the application prospects in terms of practical engineering were investigated.