The measurement error of the filling batching system is the key factor affecting the quality of the filling body. In order to reveal the effect of the fluctuation in mass fraction and pumping agent dosage caused by the measurement error on the working performance of the filling slurry, based on the cement hydration mechanism and the action principle of the pumping agent, the response surface and variance analysis methods were used to carry out mechanical properties, flow properties and rheological properties tests, and industrial test verification was carried out to determine the optimal mass fraction and pumping agent dosage. The results are concluded as follows. Firstly, when the mass fraction and pumping agent dosage fluctuated within ±1%, the main effect significantly affected the 28 day strength, viscosity, slump, expansion and average viscosity coefficient, and the interaction effect significantly affected the 28 day strength, viscosity and average viscosity coefficient. Secondly, fluctuations in mass fraction and pumping agent dosage can cause changes in the proportion of free water in the filling slurry, thereby affecting the working performance of the filling slurry. Finality, it is suggested that the fluctuation range of the mass fraction of the mine should be kept between 78% and 79%, and the fluctuation range of the pumping agent dosage should be kept between 1% and 2%. The research results can provide a theoretical basis for the construction of filling batching systems and the accuracy calibration of batching equipment.

The coarse aggregate backfill has high rigidity, poor toughness, and local energy storage accumulation, which is easy to cause safety problems. The addition of fiber can improve the toughness and ductility of the backfill and enhance its mechanical properties. Using coarse aggregate, waste rock, rod sand and river sand from Longshou Mine of Jinchuan as raw materials, the effect of fiber blending process on uniaxial compressive strength of backfill was studied. Design-Expert software was used to analyze the influences of various factors (slurry mass concentration, fiber volume rate, and cement content) on the early mechanical properties of backfill and optimize the parameters. The nonlinear regression models between the early uniaxial compressive strength value and various factors were established to reveal the interaction effects between different factors, and the cost of filling slurry material after adding fiber was calculated. The results show that the addition of fiber can significantly improve the compressive strength of the backfill. Compared with the blank group without fiber (3.03 MPa), the strength of the backfill in scheme Ⅱ (wet mixing of filling materials firstly, and then adding fibers in three stages) is 4.35 MPa, with an increase of 43.56%, which is the optimal scheme. The significant influencing factors of early mechanical properties of fiber reinforced backfill are ordered as slurry mass concentration > cement content > fiber volume rate. The interaction effect between slurry mass concentration and cement content is the most significant, which verifies the reliability of the regression model. After adding fibers, the cost of filling slurry material only increases by 0.67%−15.6%. On the premise of meeting the strength requirements of the backfill, the content of coarse aggregate and cement can be appropriately reduced, which can also reduce the cost.

With the increase of mining depth, the backfill body is affected by high ground stress, mining disturbance and groundwater erosion. In order to explore the influence of groundwater erosion on the mechanical properties of cemented backfill body with polypropylene fiber tailings, uniaxial compression test and Brazilian splitting test were carried out on specimens with different fiber contents, and specimens with better mechanical properties were selected. Based on the groundwater erosion environment, uniaxial compression and acoustic emission monitoring tests were carried out to study the damage and failure evolution characteristics of polypropylene fiber backfill body under the action of groundwater. The results show that with the increase of polypropylene fiber content, the compressive strength of backfill body increases first and then decreases, and the peak strength of the specimen with 0.3% fiber content is 3.82 MPa, which is the best. After groundwater erosion, the cumulative ringing count characteristics of acoustic emission can be divided into three stages: initial active stage, steady growth stage and rapid growth stage, and obvious damage precursor characteristics appear in the steady growth stage. The durability of the backfill body specimens with different pH erosion of groundwater is as follows: eroded backfill body specimens with pH=9> eroded backfill body specimens with pH=7> eroded backfill body specimens with pH=5> non eroded backfill body specimens. With the increase of groundwater pH, the RA-AF shear crack signal continues to decrease, the damage and failure of the backfill body changes from shear failure to tensile failure. The research results can provide reference for the improvement of mechanical properties and durability of mine backfill body.

In order to better monitor the filling morphology effect of goaf, a study on the characterization of grouting filling morphology effect with polarizability as the target was carried out based on resistivity parameters. Seven different kinds of cement were selected and prepared with five different contents of graphite powder. The variation of resistivity and polarizability of cement-based grouting filling materials was analyzed. The electrical characteristics and compressive strength of grouting filling materials under different cement types and ratios were studied, and the application test was carried out. The indoor test results indicate that with the increase of graphite powder conten, the compressive strength decreases, and the polarizability increases. When the ratio of cement to graphite is less than 10:1, the polarizability increases rapidly. In addition to sulphoaluminate cement, the content of graphite powder has little effect on the resistivity of cement-based grouting filling materials. Under the condition that the resistivity difference before and after grouting filling in goaf is not obvious, the visual characterization of filling morphology effect can be realized by polarizability (when the polarizability value of filling material is more than 3 times that of surrounding rock, the characterization effect is obvious). The research finding which is of great significance for broadening the field of grouting geophysical monitoring and improving the monitoring effect.

In order to explore the influence of physical properties of tailings on solid flux and optimize the thickening process parameters, the total tailings samples of 10 typical metal mines were selected, and the quantitative relationships between solid flux and tailings particle size or density was systematically studied. Combined with static flocculation sedimentation and dynamic thickening test data, a solid flux prediction model based on particle size-density composite parameters was established. The results show that under the condition of static flocculation sedimentation, the type and unit consumption of flocculant significantly affect the sedimentation rate and underflow concentration, and rational regulation of flocculation conditions can effectively improve sedimentation efficiency. The solid flux is significantly positively correlated with the square root of the median particle size and the density correction value of the tailings (R²≥0.94). The particle size-density composite parameters prediction model established based on nonlinear regression can accurately characterize the quantitative relationship between the physical properties of tailings and solid flux. Under dynamic thickening conditions, the feed rate of tailings slurry is linearly positively correlated with the solid flux, and the solid content of the overflow water forms a dual constraint mechanism on the flux threshold. The comparative test shows that the dynamic thickening process can increase the underflow concentration by 10%−15% compared with the static sedimentation, which fully verifies the technical advantages of the deep cone thickener in the preparation of high concentration slurry. The research results can provide theoretical basis and technical support for efficient thickening and intelligent filling of mine tailings.

In the second-step mining of open stoping with subsequent filling mining method, the blasting vibration has significant influence on the stability of the artificial pillar formed after the cement filling in the first-step stope. Based on the background of mining in the transition from open-pit to underground in Sijiaying Iron Mine, a 3D geological model and a numerical calculation model of stope were established by using the FLAC numerical simulation method. And three mining sequences of the second-step stope in the mining panel were studied, including “from one side to the other side” “from the center to both sides” and “from two sides to the center”. When the distance between the blasting hole and the backfill was 1.0 m and 1.5 m respectively, the characteristics of the effect of blasting on the stability of the artificial pillar were obtained, and the strength demand of the backfill body was inverted. The results are concluded as follows. Firstly, when the peak pressure of the blasting load is 15 MPa, with the increase in the distance from the hole to the backfill, the maximum principal stress of the backfill in the first-step stope decreases. The concentrated stress and plastic failure range in the backfill under the “from two sides to the center” mining sequence are the least, which is the most favorable to the stability of the backfill. Secondly, when the mining sequence of the second-step stope is “from two sides to the center”, the longer the distance between the blasting hole and the backfill, the smaller the plastic zone in the backfill, which is more favorable to the stability of the backfill. Meanwhile, the larger the distribution range of the shear plastic zone formed in the two-step mining ore body, the more conducive to the carving of the two-step stope. Thirdly, the strength of cemented backfill prepared by cementing powder and unclassified tailings with a ratio of1:6can meet the stability requirements of artificial pillars.

Aiming at the problem of low prediction accuracy of existing blasting vibration velocity prediction formulas in complex ground environments, a BP neural network model based on improved grey wolf optimization (I-GWO) Algorithm was proposed. The grey wolf algorithm was improved by changing the convergence factor function of the neural network to enhance optimization accuracy, initializing the wolf pack position through chaotic mapping to accelerate solution speed, and dynamically adjusting weights based on step size Euclidean distance to improve optimization efficiency. Based on the monitoring data of blasting vibration velocity at the Lilou-Wuji Iron Mine, the I-GWO-BP model was established by selecting the blast center distance, the maximum single-stage charge amount, and total charge amount as input parameters. The results show that the convergence speed and accuracy of the I-GWO-BP model are better than those of the GWO-BP model and BP model, and the optimization effect is significant. The predicted values of the I-GWO-BP model are basically within the confidence band of the measured values ±0.08 cm/s, with an average absolute percentage error of 13.84%. Its prediction performance is significantly better than other prediction methods, and its prediction accuracy is high. The research results can provide some reference for predicting the blasting vibration velocity in mines.

In order to realize the safe and efficient mining of metal mines in the surface protection area, taking the mining of inclined orebodies under slope in Paishanlou Gold Mine as the engineering background, the roof caving characteristics, stability conditions of caving arch and its control and utilization technology in the goaf of inclined orebodies were systematically studied. It was analyzed and concluded that the inclined orebodies in Paishanlou Gold Mine conform to the characteristics of the arch caving model, and the mathematical relationships between the critical caving span and caving height of the inclined orebodies under slope were established. A goaf caving process control scheme with the subarea mining model as the core was proposed, and the subarea open-stope mining with subsequent centralized filling mining technology was developed. The results of on-site practice show that after adopting this mining technology, the mining cost is decreased by 26.8 yuan/t, the production capacity is increased by 20%, the ore loss rate is decreased by 1.5%, the dilution rate is decreased by 3%, and the mined metal quantity is increased by 427.5 kg. This method not only meets the surface protection requirements of Paishanlou Gold Mine, but also achieves the goals of low-cost, safe and efficient mining. The research results can provide technical references for the mining of low-grade orebodies in surface protection areas.

In view of the characteristics that the soft powder ore rock in the deep mining underground roadway of Jinshandian Mine can not take out the complete sample on site, the direct test of compressive strength of loose soft rock mass by rebound instrument was explored. Taking the powder ore and skarn fracture zone of −425 m and −455 m horizontal tunneling roadway in the east-west mining area of Jinshandian Mine as the research object, the rebound value was tested by rebound instrument. Based on the strength formula of concrete compressive strength detected by rebound instrument in national standards and industry standards, according to the engineering geological characteristics of Jinshandian Mine, the empirical formula was optimized, and the correlation sample of rebound value and compressive strength was constructed. Using mathematical statistics method, two modified empirical formulas for strength measurement of soft powder ore rock in Jinshandian Mine were obtained by regression. The results show that the average rebound value of −425 m horizontal powder ore rock in the east-west mining area is 19.35, and the average compressive strength is 6.58 MPa. The average rebound value of −455 m horizontal soft rock fracture zone in the west mining area is 21.92, and the average compressive strength is 14.994 MPa. The results are basically consistent with the results of the inversion of the soft powder rock roadway based on the convergence value of the roadway in Jinshandian Mine. The research results provide a feasible solution for the situation that was impossible to take samples on site for indoor rock compressive stength test under poor engineering geological conditions in underground engineering.

Dynamic disasters such as rockburst caused by mining disturbance seriously restrict the development and utilization of deep mineral resources. It is of great significance to explore the propagation process of internal cracks in rocks to reveal the rock failure mechanism and disaster warning. Based on this, the spatial-temporal response characteristics of acoustic emission of granite under uniaxial loading were monitored. The single linkage clustering (SLC) method was used to construct the SLC structure of acoustic emission events. By introducing the spatial correlation length of acoustic emission events, the spatial correlation degree between acoustic emission events at different time scales was analyzed. The results show that the three-dimensional localization and energy properties of acoustic emission events can characterize the propagation and damage evolution process of microcracks in granite specimens. As the stress increases, the link length in the SLC structure gradually decreases, and the correlation within the crack cluster increases. The spatial correlation length has experienced three stages of high-level fluctuations, stable fluctuations and sudden increase. When it is close to the fracture of granite, the spatial correlation length increases sharply due to the redistribution and transmission of stress in the sample, which can be used as the early warning point of granite instability. The SLC method provides an effective method for studying the evolution process of rock crack propagation, which can provide a reference for the early warning and prevention of dynamic disasters such as rock burst.