In order to solve the difficult problem of coal wall spalling of soft seam in deep mining, the occurrence mechanism of coal wall spalling in deep mining is studied based on the limit equilibrium theory of D‐P criterion. Through the analysis of the formula of plastic zone, it can be concluded that the coal wall spalling of soft coal seam is related to mining depth, mining height, stress concentration factor, internal friction angle of coal, cohesion, and the compressive strength of coal body. Also, the change of uniaxial compressive strength of soft coal with different moisture content is studied through laboratory experiments. The experimental results show that the uniaxial compressive strength of soft coal reaches the maximum value when the moisture content is about 3.3%. The comprehensive research shows that the cohesion and strength of coal mass can be improved by increasing the working resistance of support, water injection in advanced deep‐hole coal seam of two roadways, and pre‐grouting in shallow hole of coal wall. The control measures can effectively control the coal wall spalling of soft coal seam. The research results have important guiding significance for safe and efficient mining of soft coal seam.

To study the effect of loading rate on the progressive damage and failure characteristics of coal, an ultrasonic detector and a camera were used to measure the P‐wave velocity and record the failure process of cuboid coal samples in uniaxial compression tests with five loading rates. The mechanical properties, damage process, and failure characteristics of the samples were analysed, and the mechanism of the advancing velocity of the working face in coal failure was discussed. The results show that, as the loading rate increases, the peak strength of the sample generally shows an increasing trend, but the elastic modulus changes irregularly. The sample is more prone to local failure before the peak strength. An increase in the loading rate rapidly promotes damage in the sample and accelerates the transition from internal damage to macroscopic failure, with no obvious effect on the ratio of damage threshold to corresponding peak strength. At low loading rates, the samples mainly experienced static failure; the failure form was spalling, and the failure range was wide. At high loading rates, the samples were prone to dynamic failure in the local area, manifested as the ejection of slabs and debris. A greater loading rate produced smaller and thinner slabs and a greater ejection velocity. Properly increasing the advancing velocity of the working face is conducive to reducing spalling to prevent large‐area roof fall, but it may increase the possibility of coal burst in local areas. The results of this study provide a reference for roof control and coal burst disaster prevention on the working face in deep coal mining.

The occurrence of severe mine pressure in stope is of high risk, and the mining depth is the main control factor which affecting the occurrence of mine pressure in stope. In order to study the influence law of mining depth on stope strata behavior, taking 11318E fully mechanized working face in Xieqiao Coal Mine of Huainan Mining Group as the engineering background, the stress distribution and evolution characteristics of stope surrounding rock under mining depth were studied by similar material simulation. And the evolution laws of mechanical characteristics of stope surrounding rock under different mining depths of 600 m and 1200 m were simulated. It is concluded that with the increase of mining depth, the damage range of surrounding rock on the mining field increases significantly. The initial collapse of the immediate roof above the working face is not affected by the mining depth. The periodic weighting caving step decreases with the increase of mining depth, and the height of two zones above the working face increases with the increase of mining depth. The stress transfer and evolution characteristics of the strata above the stope were revealed. The increase of front caving angle indicates that the pressure on the support of the working face is greater, indicating that the support of the working face should be strengthened with the increase of mining depth. The research results have a guiding role in the control of surrounding rock and disaster prevention of deep mine stope.

Traditional methods for geostressfield inversion face issues such as weak physical interpretability and insufficient generalization ability. This study pioneers the application of Physics-Informed Neural Network (PINN) to this problem, developing a data- and physics-driven inversion algorithm. The framework incorporates a constitutive-equation-based regularized loss function as a hard constraint during training to ensure physical consistency. To address boundary load uncertainty, two quantification approaches—Bayesian linear regression and surrogate model optimization—are proposed to establish 95% confidence intervals for boundary coefficients. Verification based on simple three-dimensional models and actual geological models of mines shows that PINN inversion achieves a mean absolute relative error as low as 0.0772%, with an error of 15.67% under sparse sampling conditions—significantly lower than the 31.07% error of the traditional Back propagation neural network. This demonstrates excellent robustness and data efficiency. In the practical engineering application of complex geological bodies, the average error of principal stress inversion is 9.35% with a minimum error of 0.137%. All inversion results fall within the permissible accuracy range of engineering, and the stress distribution conforms to basic laws, with an average error of 0.453 in the constitutive relation. Compared with BP neural network and multiple linear regression methods, it shows obvious accuracy advantages. This method provides a new solution for intelligent ground stress prediction with high accuracy, high efficiency, and strong physical interpretability, and also lays the foundation for early identification of geological disasters.