In order to reduce the cost of mine filling, the experimental study of filling binder was carried out using low-activity slag and other auxiliary materials around Timok copper-gold mine. The optimal formulation obtained by orthogonal experiment is slag : lime : caustic soda : desulfurization gypsum=81.5:10:1:7.5. The hydration mechanism of slag-based binder was studied by XRD and SEM analysis methods. Combined with the local material price and processing cost, the production cost of slag-based filling binder was calculated, and its application range was recommended. The results show that the 28 d cementation effect of the slag-based binder is slightly lower than that of the local cement, and its cost can be reduced by 30% compared with the local cement. In Timok copper-gold mine, the application of slag-based binder to the filling of sublevel open stoping with subsequent filling method can reduce the cementing cost by 13%. The annual saving of filling cost is 1.281 million euros, which has certain application value.

In the process of pipeline transportation of tailings slurry from beneficiation plant to filling preparation station, changing the feed parameters will change the spatial distribution of tailings solid particles in the pipeline, which will affect the safety and stability of slurry pipeline transportation. Taking the sulfide ore tailings transported by pipeline from a tin ore beneficiation plant in Yunnan to filling preparation station as the research object, the rheological parameters of the tailings slurry were obtained by rheological test. The ANSYS-FLUENT numerical simulation software was used to simulate and analyze the pipeline transportation of the tailings slurry. The relationship between the relative height of the bottom bed layer and the feed parameters was fitted, and a three-dimensional relationship model between the relative height of the bottom bed layer and the feed speed and volume concentration was established. The results show that the three-dimensional relationship model between the relative height of the bottom bed layer and the feed speed and volume concentration has a good fitting effect. Most of the coarse particles with a size of +120 mesh have sedimentation in the bottom bed layer. Some of the fine particles with 120−200 mesh have sedimentation at the bottom bed layer of the pipeline when the feed speed is 1 m/s, and distributed in the carrier layer when the feed speed is above 2 m/s. More fine particles of －200 mesh are distributed in the carrier layer. There is water in the boundary layer of the top wall of the pipeline and the slurry formed by some very fine particles and water, which together form the boundary layer. The change of volume concentration of tailings slurry along the radial direction can be divided into four regions. There is a gradient decreasing effect from the bottom wall of the pipeline to the carrier-bottom bed interface, it gradually increases to stability from the carrier-bottom bed interface to the stable radial volume concentration of the slurry. It remains stable after reaching the maximum value in the stable constant region of radial volume concentration, the radial volume concentration gradually decreases from the stable point to the top wall of the pipe.

In order to improve the filling efficiency and realize the continuous and stable desanding of the tailings storage thickener, combined with the actual situation of the mine and the test results of static and dynamic flocculation and sedimentation, the particle swarm optimization (PSO) algorithm was used to optimize the support vector machine (SVM) regression analysis method. The prediction model of desanding stability was established, and the prediction model was updated by the hybrid point addition method. The optimal parameters were determined by the optimization algorithm to realize the prediction of desanding stability of the tailings storage thickener. Taking Dongxinzhuang Iron Mine as an example, the optimal scheme of thickener structure parameters and desanding parameters based on PSO-SVM was completed and verified by industrial tests. The results show that the mud layer height and the combined effect of mud layer height, sand concentration and thickening time have the greatest influence on the stability of desanding. The PSO-SVM prediction model has a good accuracy in predicting the desanding stability, with a relative error of 1.36%. The research results can provide a theoretical basis for the selection of thickener shape parameters and desanding parameters, and can effectively improve the desanding stability of thickener.

Red mud as mine filling materials can solve the storage problems, but metal ions and acid ions in mine water will corrode the filling body and affect their stability. Indoor simulation tests were conducted based on the backgrounds of Bulianta Mine and Cuncaota Mine, and the impact of mine water on the compressive strength of the filling material was characterized by the erosion resistance coefficient to determine the optimal ratio of raw materials. The variation law of compressive strength of the filling body in mine water was analyzed, and the degree of erosion of the filling body by mine water was evaluated through indicators such as compressive strength loss rate and apparent failure path. Combining with testing methods such as scanning electron microscopy and energy dpectrum analysis, the micro mechanism of the action of mine water on the filling body was revealed. The results show that when the mass concentration of the slurry is 56% and the ratio of red mud, calcined kaolin, and cement is 5:4:1, the compressive strength of the specimen at 28 d is 1.98 MPa, and the initial flowability of the slurry is 260 mm, which meets the performance requirements of the filling body and achieves large-scale consumption of red mud. After soaking in sulfuric acid and sodium hydroxide solutions for 28 d, the compressive strength of the specimens decreases by 7.3% and 2.8% respectively compared to pure water immersion conditions, with corresponding corrosion resistance coefficients of 0.92 and 0.97, respectively. The microscopic test results show that the filling body has been in an acid-base environment for a long time, the Si—O—Al bond is broken, and the gel is depolymerized, leading to the gradual reduction of the compressive strength of the filling body. This research can provide a theoretical reference for the engineering application of red mud-based filling materials in mine water.

In the process of graphite mining, due to the poor stability of graphite surrounding rock, it has brought great threat to the safety production of the mine. In view of the current situation that the spatial variability of rock mass mechanical parameters was not considered in the calculation of the instability probability of graphite ore by the existing response surface method. Firstly, based on the mathematical logic of the combination of variogram and covariance function, the correlation among variables was analyzed, and a regionalized cross-correlation random field that can reflect the spatial variability of rock mass parameters was constructed. Then, combined with the failure proximity index, the limit state equation of fractured rock mass instability was constructed. Finally, the response surface method was used to analyze the failure probability of the goaf, and the quantitative evaluation of instability risk of graphite mine goaf considering parameter spatial variation structure and response surface method was realized. The results show that the instability probability of the two sides calculated by the Monte Carlo method is 17.67%, and the instability probability of the roof is 30.19%. The instability probability of the two sides calculated by the response surface method is 17.71%, and the instability probability of the roof is 29.57%. The response surface method only uses Monte Carlo method to calculate 1/4 of the time to obtain a more accurate reliability calculation result, which simplifies the calculation process and improves the calculation efficiency.

In certain mining regions of China, the presence of thick overlying strata above the coal seam roof not only triggers rock burst disasters but also leads to significant seismic events. The occurrence mechanism of rock burst induced by thick overlying strata was analyzed. A strategy for controlling rock burst through grouting filling of the separated layer in the overlying strata was proposed. Numerical simulations and theoretical analysis were combined to investigate the load reduction and impact attenuation effects of grouting filling. The research findings indicate that in mines with thick overlying strata, the stress concentration and dynamic load intensity resulting from hanging wall effects are significantly great due to the large thickness, high strength, and considerable limit span of the rock layers. This makes them susceptible to rock burst disasters. The grouting filling technique, by filling the separated space, can support the thick overlying strata together with the surrounding rock mass, thereby reducing the sinking of the thick overlying strata. This approach effectively controls the fracture and movement of thick overlying strata, consequently reducing both dynamic and static loads and achieving the objective of minimizing dynamic overburden pressure. Numerical simulation results of grouting filling indicate that, compared to without grouting filling, the stress concentration coefficient around the separated space decreases by 8.89%,18.59%, and 35.96% when the grouting filling rates are 30%, 60%, and 90%, respectively. This decrease is associated with a substantial reduction in the height of the plastic zone and the degree of damage to the thick overlying strata, indicating that the grouting filling of separated layer can achieve the purpose of reducing load and preventing rock burst.

The anchoring operation of coal mine excavation roadway is the key factor affecting the tunneling efficiency. Accurately determining the position of anchor hole under the condition of roadway with many interference factors is one of the key links to improve the anchoring speed. At present, most of the research on anchor hole positioning has a single consideration of interference factors, and the influence of complex factors on anchor hole identification under multi-scale changes has not been studied. To this, an improved YOLOv8s anchor hole positioning method in multi-scale complex environment was proposed. Firstly, according to the actual anchoring situation of the roadway, four main factors affecting the positioning of the anchor hole, such as light intensity, dust concentration, water mist concentration and shooting distance, were summarized, and a three-level influence scale was established. The initial data set was formed by the combination of various influencing factors. Before the anchor hole identification, the image preprocessing operation was performed to optimize the overall quality of the image. Then, the YOLOv8s deep learning network model was improved, and the multi-scale hole attention mechanism (MSDA) was added to improve the feature extraction ability of the model under complex conditions. Finally, the pixel coordinates of the anchor hole center were calculated to obtain the actual three-dimensional coordinates. The results show that the average detection precision of the improved YOLOv8s model reaches 91.6%, which is 5.3 percentage points higher than that of the YOLOv8s model. The improved YOLOv8s model has better detection ability and can accurately detect the anchor hole position in the image under the interference of different scale combination influencing factors. The number of frames per second of the processed image of the improved YOLOv8s model is maintained at 171 frames, which can meet the real-time requirements of the model detection function.

In order to study the damage effect of overlying strata and the movement and instability characteristics of overlying strata under mining in close-distance coal seams, based on the mining of a close-distance coal seam in Shendong mining area, the damage theory was used to analyze the instability and deformation characteristics of the overlying strata of the close-distance coal seam under mining and the time-space evolution law of microseismic. The damage characteristics of the close-distance coal seam considering the microseismic monitoring data were discussed. The 3DEC software was used to simulate the instability and fracture of overlying strata and the law of stress transfer in close-distance coal seams under mining. The results show that the damage evolution of overlying strata increases as the distance between the upper and lower coal seams approaches. The minimum damage degree is 0.31 and the maximum damage degree is 0.77 under different monitoring schemes. There is no obvious increase or decrease trend of microseismic events in the process of overhaul, and the number of microseismic events increases continuously as the closing distance between the upper and lower coal seam working faces during coal mining. The continuous advancement of the working face of the lower coal seam causes the displacement of the overlying strata to increase, and the maximum displacement of the overlying strata is 45.9 cm. The research results are of great significance for the stability control of overlying strata in similar close-distance coal seam mining and the safe and efficient mining of coal mines.

In order to reduce the adverse effect of creep on the long-term stability of the project under the action of seepage water pressure, creep tests under different hydraulic gradients were carried out using the seepage-stress coupling servo control loading test system to explore the rheological properties and seepage characteristics of sandstone. Combined with three-dimensional digital image correlation technology (3D-DIC), the micro-cracks evolution law and deformation localization characteristics of sandstone were studied. At the same time, the crack damage characteristics of sandstone were explored based on CT scanning technology. The results are concluded as follows. Firstly, the larger the hydraulic gradient is, the larger the creep deformation is, the faster the micro-cracks growth rate is, and the more obvious the change of strain field cloud map is. Compared with the axial strain field cloud maps, the larger strain points and strain localization characteristics appear earlier in the radial strain field cloud maps. Secondly, the larger the hydraulic gradient is, the larger the strain increment is, the faster the strain change rate is, and the shorter the creep life of sandstone is. By establishing a virtual strain gauge, it is found that the strain changes inside and outside the local failure zone of sandstone are inconsistent, indicating that there are localization characteristics in the creep deformation process of sandstone. Thirdly, under the low hydraulic gradient, more pores and fissures are pressed and sealed, the seepage channel is reduced, and the flow rate is continuously reduced. Under the high hydraulic gradient, the damage is continuously accumulated, the seepage channel is increased, and the flow rate is increased. Finally, it can be seen from the CT scanning that the three main cracks penetrate the whole sample, guide the sandstone failure, and generate secondary cracks to promote crack propagation. Near the loading end of sandstone, the applied seepage water pressure is larger, and more secondary cracks are derived from the upper part of sandstone, forming a complex crack network structure.

The rock mass strength is a key mechanical parameter for the safety and stability analysis of mining engineering. Combined with the key project of small-scale mechanized mining technology in Shanhu Tungsten Mine, based on the detailed investigation of rock mass structural plane, three methods of BQ, Q and RMR were used for classification of rock mass quality and stability evaluation. Then, the rock mass strength was estimated based on quantitative GSI and Hoek-Brown Criteria to obtain accurate and reliable parameters of rock mass strength. The research results are concluded as follows. Firstly, the dominant joint plane strikes NNW-SSE or NW-SE, inclines NE or SW, and the dip angle is large. The volume joint number J_v is 7.07−9.88/m³, the integrity coefficient K_v is 0.55−0.63, and the integrity evaluation is relatively intact. Secondly, the [BQ] value, Q value, and RMR₈₉value of the limestone rock mass are 380.96, 5.58, and 59, respectively, and the grade of rock mass quality is Ⅲ or average. Thirdly, according to the BQ classification results, when the span of roadway or stope is less than 5 m, the surrounding rock can be basically stable. According to the Q classification results, the maximum unsupported span of roadway or stope is about 6.35 m. According to the RMR classification results, when the span of roadway or stope is 5 m, the average standing stability time of surrounding rock can reach 7 days. Finally, based on quantitative GSI and Hoek-Brown criteria, the Hoek-Brown strength parameters and equivalent Mohr-Coulomb strength parameters of limestone rock mass are obtained, which can provide accurate and reliable mechanical parameters for mining design and the stability analysis of rock.