Due to the significant differences in the intensity characteristics of near-fault ground motion in different directions, taking typical near-fault seismic records as an example, three ground motion intensity parameters were selected to compare the variability of ground motion intensity in various directions. Based on OpenSees finite element software, the difference of seismic response of slopes under different directions of seismic was analyzed. Considering the strain softening characteristics of soil, the influence of directional effects of seismic motion on the seismic stability of slopes was explored. The results show that the PGA variability of peak acceleration in different directions reaches 62.5%. The displacement response of the slope top in the direction of the maximum PGA under seismic action increases by 20.5% and 86.5% respectively compared to the two components recorded in the measured horizontal seismic motion. The variation range of slope safety factor under different directions of seismic action is 0.93-1.15. Ignoring the seismic direction may overestimate the seismic capacity of the slope and increase the risk of seismic instability. The research results can provide reference for slope seismic design and disaster prevention.

To reveal the loading rate effect of roof fracture behavior, typical mine roof sandstone was taken as the research object. Brazilian splitting tests of five loading rate levels were carried out in the loading rate range of 0.001-10 kN/s. The mechanical behaviors such as tensile strength, fracture characteristics, initiation stress, acoustic emission characteristics and fracture microstructure were analyzed. Based on the distribution of acoustic emission sources, the evolution law of internal damage cracks in the samples under different loading rates was explored. The results show that the tensile strength increases by 0.64 MPa when the loading rate increases by one order of magnitude, but the ratio of crack initiation stress to tensile strength decreases. With the increase of loading rate, this ratio is 40.13%, 36.59%, 32.53%, 26.17% and 15.81% in sequence. When the loading rate increases by one order of magnitude, the average size of the fragments decreases by 0.229 mm, the degree of fragmentation increases, and the size distribution of the fragments develops towards homogenization. The acoustic emission source at low loading rate mainly appears along the macroscopic fracture surface, while at high loading rate, the distribution of acoustic emission source is more dispersed due to the initiation and propagation of cracks from multiple dispersed stress concentration areas. With the increase of loading rate, the transgranular phenomenon at the fracture surface becomes more and more significant, transitioning from a single crack and straight form at low loading rates to multiple cracks and complex forms at high loading rates.

To investigate the mechanical properties and energy evolution laws of fractured sandstone under water-bearing conditions, uniaxial compression tests were conducted using a microcomputer-controlled coal-rock rigid impact testing system on sandstone specimens collected from a mine in Ordos. The tests considered both natural and saturated states, as well as different fracture inclination angles (0°, 30°, 60°, 90°). The full stress-strain curves, physical and mechanical parameters, energy evolution laws, and crack propagation behavior of fractured sandstone were analyzed. The results show that compared with intact sandstone, the peak strength and elastic modulus of sandstone with pre-existing fractures are significantly reduced. The total energy absorbed by the specimens under natural states (196-317 kJ/m³) is approximately 1.85 times that under saturated states (102-176 kJ/m³), and the total energy increases as the fracture inclination angle increases. Regarding energy evolution, the proportion of elastic strain energy exceeds 92% during the stable damage stage, while it drops to below 61.5% during the accelerated damage stage, with the proportion of dissipated energy which shows the opposite trend. The failure mode of the specimens exhibits significant anisotropy, and the deterioration effect of water reduces the initiation stress by 13.43% to 18.18%, while also increasing the occurrence of rock spalling. Fracture inclination angle and water-bearing state are key factors controlling the mechanical properties, energy evolution, and failure mode of sandstone. The research findings can provide a theoretical basis for the stability control of water-rich fractured rock mass engineering.

To quantitatively analyze the dynamic response characteristics of prefabricated fractured rock mass under the coupled action of impact load and confining pressure, a SHPB numerical model was established using ANSYS/LS-DYNA based on one-dimensional stress wave theory and the HJC dynamic damage constitutive model. Dynamic impact simulations were carried out on prismatic marble specimens with different prefabricated fracture inclination angles and different initial confining pressures. The results show that under a given strain rate, the pre-peak segments of the stress-strain curves for various working conditions are basically coincident, and macroscopic cracking mainly occurs in the short duration near the peak stress. Confining pressure has a significant enhancement effect on the dynamic strength. The peak stress increases with the increase of confining pressure, and its increase with the variation of fracture inclination angle is more significant under high confining pressure. When the confining pressure is 0, 5 and 10 MPa, the increase of peak stress with the variation of fracture inclination angle is approximately 14%, 20% and 30%, respectively. The peak strain of the rock mass increases from about 0.004 to about 0.006 1. The crack morphology is generally dominated by "X"-type shear failure. Based on the crack propagation trajectory and the crack initiation mechanism, five typical crack types and one fragmentation characteristic can be summarized. As the confining pressure increases, the crack network inside the rock mass becomes more complex and the fragmentation degree of the specimen intensifies significantly. Energy evolution analysis shows that when the incident energy is kept consistent across all working conditions, as the confining pressure increases from 0 to 10 MPa, the average energy absorption rate of the specimen increases from 6.13% to 9.35%, the energy reflection rate decreases from 93.54% to 89.99%, and the energy transmission rate increases from 0.33% to 0.66%. The dissipated energy generally increases with the increase of confining pressure. As the fracture inclination angle increases, the dissipated energy shows a local low value near 45° and then increases rapidly. The research results can provide a reference for damage assessment and stability control of fractured rock mass under the coupled action of blasting vibration disturbance and in-situ stress in scenarios such as shallow slopes and deep roadways.

The strength prediction of grouting materials is of great significance to improve the reliability of material design for grouting. Aiming at the problems that the traditional empirical models are difficult to accurately capture the relationship between material composition and strength, and the insufficient generalization ability of models caused by small samples, a machine learning prediction framework combining physical mechanism and data enhancement was proposed. The material performance prediction results of four machine learning methods, Random Forest Regression (RFR), Artificial Neural Network (ANN), Support Vector Regression (SVR) and Linear Regression (LR), were compared and analyzed. The results show that the RFR model has the best predictive performance, and its determination coefficient (R²) is increased to 0.934, which is 21.8%, 10.4% and 1.6% higher than that of LR, SVR and ANN models, respectively. The mean square error MAE is reduced to 2.59, which is 14.5%, 13.4% and 9.1% lower than that of LR, SVR and ANN models, respectively. In the analysis of feature importance, the contribution of graphite ratio to compressive strength is 37.6%. The data enhancement technique based on physical constraints can significantly improve the robustness of small sample modeling. The MAE of RFR and ANN models increases by 39.8% and 18.3% respectively compared to those without data enhancement. The prediction error of the RFR model remains stable below 10% under various water-solid ratio conditions. The research results can provide reference for the proportion of grouting reinforcement materials in mines.

Grouting materials have significant applications in mining engineering, but their poor durability in deep and highly erosive environments severely restricts the long-term safe operation of mines. Modified grouting materials were prepared by replacing a portion of the cement with rice husk ash and slag. Using compressive strength, the strength erosion coefficient, and the mass loss rate as evaluation criteria, quantified assessments were conducted to evaluate the impact of different erosion environments on the durability of modified grouting materials in karst regions. Combining scanning electron microscopy and X-ray diffraction analysis, it revealed the erosion-resistant mechanism of the modified grouting material at the microscopic level. The results show that the rice husk ash-slag-cement grouting material exhibited the highest resistance to chloride-sulfate erosion, carbonate erosion and water leaching corrosion, with strength loss rates of 13.61%, 14.42%, and 4.79%, respectively, and mass loss rates of 3.30%, 2.38%, and 1.95%, respectively. It is more resistant to environmental erosion than pure cement grouting materials. At the microscopic level, it is observed that rice husk ash and slag form dense hydration products through hydration reactions, which can inhibit the degradation of the original hydration products caused by various corrosive environments. At the same time, the amorphous SiO₂ in rice husk ash reacts with Ca(OH)₂ through a pozzolanic reaction, and the secondary hydration of the slag generates more C-S-H gel, significantly enhancing the erosion resistance of the modified grouting material to chloride-sulfate, carbonate and water. The cost of grouting materials made from rice husk ash-slag-cement is 21% lower than that of pure cement, and they can extend the maintenance cycle of the project by approximately 51.41%.

Safety protection for hoist cars in ordinary inclined shafts primarily relies on passive safety devices installed within the shaft. To prevent catastrophic consequences resulting from an accelerated impact, a structure for an inclined shaft hoisting container was proposed, featuring mechanical braking, maintenance interlocks, and spring-loaded shock absorbers. By establishing a multi-body coupled impact dynamics model of the hoisting container-slope track system, the formation mechanism and key influencing factors of the rear wheel lift angle during braking impacts were revealed. Combining theoretical derivations, virtual prototyping simulations, and experimental validation, the effects of buffer spring stiffness, slope track angle, impact velocity, and various operational loads on the dynamic behavior of the hoisting container were systematically analyzed. The results show that the system impact force and rear wheel lift angle reach the minimum values when the spring stiffness is approximately 2 500 N/mm and 15 000 N/mm, respectively, representing reductions of 13.4% and 17.4% compared to the springless condition. The slope track angle is positively correlated with impact velocity and the rear wheel lift angle. The rear wheel lift angle under rated load conditions is 15% to 30% lower than that under no-load conditions. When the slope track angle increases from 10° to 30°, the derailment impact speed under rated load conditions decreases from 9.5 m/s to 7.3 m/s, and under no-load conditions, it decreases from 9.0 m/s to 5.2 m/s, resulting in a maximum safety margin of 28.8% for load-carrying capacity. The research findings provide a theoretical basis for improving container impact-resistant design and safe braking strategies.

Shotcrete support is widely used in roadway excavation support, with the advantages of low cost and high safety. The robot arm is the core part of the shotcrete robot. In order to improve the design theory, kinematics and trajectory planning analysis were carried out taking a six-degree-of-freedom robot arm as the research object. For the time-optimal trajectory planning of the robot arm, a trajectory planning method based on the improved whale optimization algorithm was proposed. By introducing the Circle chaos mapping, Cauchy perturbation strategy and the pheromone mechanism of the ant colony algorithm, the global search ability of the algorithm was enhanced, which effectively avoided being trapped in the local optimal solution, and significantly accelerated the convergence speed, while improving the efficiency and accuracy of the local search process. The improved whale optimization algorithm enables the robot arm to reach the target posture more quickly, successfully shortening the running time by 56.7%, thereby further verifying the feasibility of the algorithm.

To investigate the collision characteristics of grinding balls inside the cylinder of a vertical stirred mill during operation, a laboratory-scale vertical stirred mill was used as the research subject. Based on the Discrete Element Method (DEM) and Computational Fluid Dynamics (CFD), a DEM-CFD simulation model was developed, taking grinding balls, slurry, and the agitator as the objects. The collision characteristics of grinding balls and their variation along the axial direction were systematically investigated based on five aspects, including the number of collisions, normal and tangential collision energies, and normal and tangential collision forces. The results show that collisions among grinding balls account for 90.5% of the total number of collisions, which is significantly higher than the number of collisions between grinding balls and the grinding cylinder. The tangential collision energy among grinding balls is approximately 1.4 times that of the normal collision energy, while the normal collision force is approximately 6.2 times that of the tangential collision force, indicating that the grinding effect of the grinding balls is more pronounced in the tangential direction. Grinding balls located between the bottom and the 3/8 mark of the cylinder exhibit collision characteristics marked by high collision frequency, high collision energy, and low collision force, which effectively refine the materials. The grinding effect in other areas is relatively small, suitable for rapid materials grading and allowing materials that meet the particle size conditions to overflow from the cylinder. This research achievement provides theoretical guidance for optimizing mill design and improving grinding efficiency.

To address the issue of insufficient surface feature recognition accuracy in traditional segmentation models caused by complex vegetation cover and blurred surface feature boundaries in mining remote sensing images, a dual-branch DB-TransUNet model was proposed. This model innovatively constructed a dual-branch encoding architecture and integrated features from both branches using an adaptive weighted fusion (AWF) module. By combining this with a channel adaptive enhancement (CA) module to optimize feature weight allocation during the decoding process, it achieved focused detection of features in key regions and effective suppression of background noise. The experimental results show that incorporating the AWF and CA modules improved the mIoU of the model by 4.16% and 78.63%, respectively. The full DB-TransUNet model achieves excellent segmentation results of mIoU=0.903 7 and mDice=0.948 8 in complex remote sensing scenarios. Compared to mainstream network segmentation architectures, the DB-TransUNet model demonstrates superior performance across all indicators, achieving higher accuracy and stronger generalization capabilities in the recognition and detailed segmentation of complex surface features in mining environments. The findings validates the effectiveness of a strategy combining complementary multi-source features with weighted fusion in mitigating vegetation-induced interference in the identification of mine surface features, thereby providing a reliable technical solution for high-precision intelligent monitoring in open-pit mining areas.