Aiming at the problem of the safe thickness of the isolation layer in an iron mine from open-pit to underground mining, the technical means combining theoretical calculation and numerical simulation was adopted to analyze the influence characteristics of the underground ore body mining process on the surface settlement and rock movement under the condition of different thickness isolation layers. The results show that the safe thickness of the isolation layer of the iron ore is from 11.25 m to 24.41 m based on the theoretical calculation method. The numerical simulation results show that when the thickness of the isolation layer is 15 m, 20 m and 25 m, the surface settlement displacement of the surface structures, the open-pit slopes and the dump sites are all small, which can meet the stability evaluation standard of the rock mass. Considering the safety of mining and avoiding the loss of mineral resources, the safety thickness of the isolation layer is recommended as 20 m. After the implementation of the plan, it is expected that the comprehensive production capacity of the mine will increase by 20%, the safety accident rate will decrease by 53%, it has advantages in terms of safety reliability, technical feasibility and economic rationality. The study can provides an important reference for similar mines to conduct open-pit to underground mining.

In order to reduce the cost of filling mining, slag powder-cement was used as a composite cementitious material, and the response surface method was used to obtain the optimal ratio of filling materials that could meet the strength requirements of a mine filling body. Firstly, through the strength test of filling body with different cement-sand ratios (slag powder-cement : tailings) and the slump test with different slurry mass concentrations, the ranges of slag powder dosage and slurry mass concentration were preliminarily determined. Then, based on RSM-BBD module, the strength regression model of slag-based cemented filling body at different ages was established, and the influences of slag powder dosage, cement-sand ratio and slurry mass concentration on the strength of cemented filling body were analyzed. The results show that the interactions of slag powder dosage and slurry mass concentration, slag powder dosage and cement-sand ratio, cement-sand ratio and slurry mass concentration have significant effects on the 3 d strength, 7 d strength and 28 d strength of the filling body, respectively. Finally, combined with the engineering requirements, the best ratio of filling materials was recommended. Mine practice shows that the use of slag-based cementitious materials instead of cement for mine filling can reduce the cost of filling cementitious materials by 12%−13%.

In order to study the strength formation mechanism of sulphoaluminate-based grouting backfill with different additions of fly ash under acidic environments, the early strength development and microstructure evolution of sulphoaluminate-based grouting backfills were characterized by uniaxial compression tests, XRD phase analysis, thermal analysis (TG) and MIP mercury porosimeter pore tests. The results show that when the addition of fly ash in the grouting backfill is 10%, the strength at 1 d of hydration is 11.05 MPa, and the strength at 7 d of hydration reaches 26.73 MPa. When the addition of fly ash is increased to 50%, the strength at 1 d of hydration is only 4.06 MPa, while the strength at 7 d of hydration is 10.05 MPa, which indicates that the strength loss caused by high addition of fly ash is obvious. Under the influence of acidic environment, the strength of grouting backfill of each ratio decreases. The strength loss of sample adding 10% fly ash is 17.60%, 19.36% and 21.42% respectively, and the strength of sample adding 50% fly ash is decreased by 21.01%, 28.89% and 39.45% respectively. High fly ash addition and acidic environment can increase the threshold pore size and porosity of the backfill, resulting in a decrease in the early strength of the backfill. In the actual grouting reinforcement, the negative effects of acidic environment and external additives on the setting rate of the cementitious system and the early strength of the backfill should be fully considered to ensure the grouting reinforcement effect of the coal roadway.

The blasting quality of fan-shaped blasthole has a significant effect on the application effect of non-pillar sublevel caving method. In view of the problems such as serious damage of eyebrow line, uneven blasting fragmentation and large loss and dilution when the non-pillar sublevel caving method was used in the west second mining area of Longshou Mine under the condition of ore-rock breaking, ANSYS / LS-DYNA was used to optimize the charge structure parameters of the fan-shaped blasthole. The results show that the blasting effect of fan-shaped blasthole can be effectively improved using W-type staggered charge at orifice, hole bottom initiation and millisecond blasting in row. The optimized blasthole charge structure parameters are applied to the field production practice. The ore recovery rate is increased by about 36%, the ore dilution rate is reduced by about 31%, the bulk rate is reduced by 52%, and the damage of the eyebrow line is also significantly reduced.

Aiming at the problems of poor durability and water stability of traditional gravel transportation roads in open-pit mines, the transportation roads were modified using solid waste around the mine. The ratio of blasting waste rock and surrounding soil in the cement stabilized layer was determined by laboratory tests. The unconfined compressive strength, water stability and durability of the cement stabilized layer with different the dosages of curing agent were discussed. The results show that the unconfined compressive strength of the unmodified material is lower than 0.2 MPa, and when the ratio of blasting waste rock and surrounding soil is 4:6, the strength of blasting waste rock can be fully utilized and the pavement can be well formed. The curing agent can significantly improve the strength of the material. When the dosage of curing agent is 8%, the unconfined compressive strength exceeds 2.5 MPa, the water stability coefficient is 137.68%, and the saturated compressive strength of the sample exceeds the initial unconfined compressive strength. In the dry-wet cycle test, the compressive strength of the sample increases significantly and then decreases slightly with the increase of the number of cycles. When the dosage of curing agent is more than 6%, the durability is significantly improved. The deflection value test and comprehensive cost analysis were carried out using the combination of stone-soil ratio of 4:6 and dosage of 8%. The results show that the deflection value of the modified pavement is 9.07 (0.01 mm), the pavement quality was significantly improved, and the comprehensive cost is 47.5 yuan/m², which is significantly lower than that of the original gravel pavement.

In response to the problems that the traditional prediction model is easy to fall into local optimal solution, the weight and threshold of BP neural network were optimized by seagull optimization algorithm (SOA), and the SOA-BP prediction model was constructed. Based on the time series settlement values of multiple high coherence points, the model can accurately predict the surface settlement in a large-scale in the mining area. The SBAS-InSAR technology was used to process 59 SAR images of the research area, and the time series settlement values of 6 997 high coherence points in the area were extracted. Based on the monitoring data, the SOA-BP prediction model was used to predict the surface settlement in the research area, and the prediction accuracy was calculated. The results show that compared with BP model, the maximum root mean square error of SOA-BP model is reduced by 82.77%, and the maximum mean absolute error is reduced by 81.13%. With the passage of time, the proportion of samples with residual distribution within ±5 mm of SOA-BP model is increased from 79.59% to 91.91%, and the maximum residual extreme value is decreased from 16.722 4 mm to 9.421 9 mm. The prediction accuracy of SOA-BP model is high, and it can accurately predict the law of surface settlement.

Aiming at the limitations of traditional methods for determining the rock movement angle under the geological conditions of steeply inclined multiple ore veins, a comprehensive analysis method based on the critical granular column theory and numerical simulation was proposed. Taking the Shagou stope of Faende mining area in Henan Province as the research object, through engineering analogy, theoretical analysis and numerical simulation, the influence law of the mining of steeply inclined multiple ore veins on the rock movement angle was deeply studied. The nonlinear relationships between rock movement angle and vein distance, vein quantity, as well as dip angle of ore body were revealed, and corresponding empirical formulas were established. The results show that the range of the rock movement angle in the mining area is from 69° to 78°, and the proportion of the height of the critical granular column ranges from 15.8% to 20.7%. By combining the height of the critical granular column with the dynamic correction model of the rock movement angle, the delineation scheme of the safety pillar was optimized, increasing the resource utilization rate to 82.3%. Based on the results of theoretical calculation and engineering practice, a dynamic correction model of the rock movement angle under the coupled action of multiple ore veins was proposed, breaking through the applicable boundary of the traditional analog method for single ore veins, and providing a strong basis for the safe mining of similar complex steeply inclined multiple ore veins.

In response to the problems of imbalanced sample data and low prediction accuracy in the construction of rock burst intensity grade prediction models, a rock burst intensity grade prediction model (BO-CNN-LSTM-Multi-head Attention) based on convolutional neural networks (CNN) and long short-term memory networks (LSTM) was proposed by integrating Bayesian parameter optimization algorithm (BO) and Multi-head attention mechanism (Multi-head Attention) under the rock burst sample data balancing strategy. 224 sets of rock burst cases were collected from both domestic and international sources, and Box-Cox transformation and support vector machine synthesis for few class oversampling technique (SVM SMOTE) were used to normalize and balance the rock burst sample data. The processed dataset was input into the CNN-LSTM prediction model for training and prediction, and the model was gradually optimized by Bayesian parameter optimization algorithm and multi-head attention mechanism. The prediction results indicate that Box-Cox transformation and SVM SMOTE sampling can effectively reduce the impact of outliers and data imbalance on the model, and enhance the separability of the dataset. The accuracy of the BO-CNN-LSTM-Multi-head Attention rock burst intensity grade prediction model is 95%. This model can effectively improve the prediction accuracy of rock burst intensity in the case of imbalanced sample data, and has a high feasibility and practicality.

The layered characteristics of rock mass in the main vein roadway of a lead-zinc mine in Yunnan with a buried depth of 1 500 m are obvious, the ground pressure is severe, and the stability control of surrounding rock mass is difficult. Based on the engineering geological investigation and test materials, the numerical model of engineering geological body was established using 3DEC numerical simulation software. The deformation and failure characteristics of surrounding rocks of roadway were compared under two working conditions of unsupported and designed support after excavation of roadway with different dip angles of the rock strata. Then, the actual deformation monitoring of roadway was carried out. It is found that the horizontal and vertical displacements of the roadway are significantly affected by the dip angle of the rock stratum. As the dip angle of the rock stratum increases, the horizontal displacement gradually increases and the vertical displacement gradually decreases. The deformation of surrounding rocks decreases obviously after roadway support, and the shape of plastic zone is similar to that of unsupported roadway, but the range is smaller. After applying the support, the roadway roof displacement monitoring was carried out. It is found that with the passage of time, the roof displacement gradually increases and tends to be stable after about 70 days, and the displacement of the roof gradually decreases with the increase of the dip angle of the rock stratum. Therefore, the support system used in mines can effectively reduce the deformation of roadway in deep layered rock mass, especially the horizontal deformation of the roadway vault and the vertical deformation of the right side, which is conducive to maintaining the long-term stability of the roadway.

When there is a cavity in the surrounding rock of the roadway, it will have a serious adverse effect on the stability of the roadway. In order to solve the problem of significant deformation and failure caused by the influence of filling unconnected roof cavity under the floor in the last sublevel stoping route of caving method stope in Longshou Mine of Jinchuan, the failure mechanism and support mode of the mining approach were studied by means of field investigation, numerical simulation and field test. The results show that the unconnected roof cavity is deformed under the action of stress concentration after the excavation of the approach, which leads to the synergistic deformation of the surrounding rock of the approach, and the existence of the cavity provides the development space for the deformation of the surrounding rock, which is the main reason for the serious deformation and failure of the approach. According to the position of the cavity and its control effect on the stability of the mining approach, an asymmetric support scheme was proposed. In this scheme, the unconnected roof cavity was regarded as the weak point of the whole support system to impose support, so as to reduce the adverse effect of the cavity on the stability of the approach. The field application results show that compared with the original support scheme, the asymmetric support scheme significantly improves the support effect, reduces the deformation of the two sides of the approach by more than 80%, and reduces the support cost by about 11.7%, which improves the stability of the stope and ensures the safe production of the stope.