With the widespread application of lithium-ion batteries in underground mines, the safety issue of mining batteries has become increasingly prominent. The thermal runaway characteristics of large capacity lithium iron phosphate batteries for mining were studied by overcharging tests of 200 Ah LiFePO₄/C battery cell and battery module under different overcharging rates (0.5 C, 1 C, 1.5 C). The results show that the thermal runaway behaviors of the lithium iron phosphate batteries are divided into three stages: shell expansion, slow flue gas injection, and violent flue gas injection with subsequent natural cooling. As overcharging rate increases, the overcharged capacity required in each stage gradually decreases. The temperature after thermal runaway of the battery can reach up to more than 400 ℃, and the maximum temperature in the battery module test is significantly higher than that in the battery cell test. High temperature will pose a severe challenge to the safety of underground mines, and corresponding cooling and protective measures need to be taken. The thermal runaway effect of overcharged battery in the battery module does not cause thermal runaway reactions of adjacent batteries. The critical conditions for the thermal runaway chain reaction of mining batteries still need further study.

Based on the first-principles of density functional theory, quantum chemical calculations were performed using the CASTEP module of Materials Studio (MS) software to simulate and optimize the crystal structure, cleavage surface, and adsorption models of magnesite and hornblende in adsorption with reagents. On this basis, the band structure and density of states of magnesite and hornblende were analyzed. The adsorption energies of dodecylamine and the novel collector KDLX on the magnesite (104) and amphibole (110) surfaces were obtained, respectively. The results show that the band gap widths of magnesite and hornblende are 4.920 eV and 3.962 eV. The optimized crystal structure has a better stability. The ammonium hydrogen atoms in dodecylamine and KDLX undergo hydrogen bonding and physical adsorption with the oxygen atoms of minerals. Compared with dodecylamine, KDLX has a stronger adsorption capacity for hornblende and it is predicted that the collector can be used for flotation removal of silica-containing gangue minerals in magnesite ore. This study has revealed the surface characteristics of minerals and the adsorption mechanism of reagents, which has a guiding significance for the flotation separation of magnesite and hornblende and the selection of flotation reagents.

China's iron ore resources are increasingly depleted, and they have difficult to select characteristics such as poor, fine and miscellaneous. The traditional anionic collector has large dosage, complex reagent system and poor activity. In order to solve this problem, the surfactant processed by industrial waste amine (YTDB) was used as cationic collector to study the single mineral flotation test and mineral adsorption mechanism of quartz and hematite. The results show that when dodecylamine (DDA) is used as collector, under the conditions of pH=7 and collector dosage of 20 mg/L, the recovery rate of quartz is 78.36%, and the recovery rate of hematite is 12.57%. At this time, the flotation difference is the largest. When YTDB is used as collector, under the condition of pH=7 and collector dosage of 15 mg/L, the recovery rate of quartz is 91.27%, and the recovery rate of hematite is 12.67%. At this time, the flotation difference is the largest. YTDB is obviously better than DDA in flotation index, and saves the dosage of reagent to a certain extent. By testing the infrared spectrum, surface tension and Zeta potential before and after the interaction of the agent and the mineral, it is found that YTDB adsorbed on the surface of quartz.

The contents of silver (Ag), lead (Pb), zinc (Zn), iron (Fe), manganese (Mn) and sulfur (S) in a polymetallic ore in Heilongjiang are 330.52 g/t, 0.57%, 0.29%, 25.77%, 9.05% and 3.38%, respectively. Due to the carbon content is high, mineral dissemination relationship is complex, as well as Ag, Pb and Zn are partially wrapped in Fe-Mn minerals, it is difficult to guarantee the concentrate grade and recovery rate. In order to realize the comprehensive recovery of valuable components in the ore, the process of Ag−Pb−Zn mixed flotation and Fe-Mn magnetic separation was adopted, and the full process closed-circuit test was carried out. The Ag, Pb and Zn grades of the Ag−Pb−Zn mixed concentrate obtained by the test are 3 010.80 g/t, 5.39% and 2.37%, respectively, and the recovery rates are 83.35%, 89.61% and 77.55%, respectively. The Fe grade and recovery rate of the Fe concentrate are 60.97% and 8.01%, respectively, and the total Fe recovery rate is 71.89%. The Mn grade and recovery rate of Mn concentrate are 18.30% and 76.38%, respectively. The tailings yield of the full process is 50.23%, and the Ag, Pb and Zn grades of the tailings are 35.07 g/t, 0.055% and 0.068%, respectively. By using copper sulfate as an activator and butyl xanthate+ammonium dibutyl dithiophosphate as a collectors in Ag−Pb−Zn mixed flotation, the process has realized the comprehensive recovery of Ag, Pb and Zn with high recovery rates, as well as the effective recovery of Fe and Mn, additionally, the backwater can be reused in production.

In order to promote the leaching of Ca and Mg elements in anorthite, a strain hzt-1′ with good adaptability to anorthite was screened from the soil of Jialiangshan, Fangshan County, Lvliang City, Shanxi Province. It was identified as Phyllobacterium myrsinacearum. The optimum growth conditions of the strain and the optimum leaching conditions of anorthite in Czapek Medium were investigated. The optimum growth conditions of strain hzt-1′ are as follows: pH of 6, liquid volume of 80 mL, inoculation volume of 7%, glucose as the best carbon source (20 g/L), and NaNO₃ as the best nitrogen source (1 g/L). The optimum leaching conditions of strain hzt-1′ are as follows: pH of 6, inoculation volume of 3%, pulp mass concentration of 20 g/L, and liquid volume of 100 mL. Under these conditions, the leaching rates of Ca²⁺ and Mg²⁺ in anorthite are 50.3% and 39.91%, respectively. The strain promotes the dissolution of anorthite through proton exchange and complexation, accelerates the leaching of Ca²⁺ and Mg²⁺, and thus improves the utilization rate of anorthite and other mineral resources. This study can provide a new strain for efficient leaching of silicate minerals.

As an environmentally friendly material, gold-tailings-based concrete has a wide range of potential applications. However, the complexity of the material composition of gold-tailings-based concrete, traditional prediction methods of compressive strength are often difficult to capture the nonlinear correlation and multivariate coupling characteristics within the material, resulting in insufficient prediction accuracy. Thus, a strength prediction model for gold-tailings-based concrete was proposed based on a deep learning binary fusion model (DP), a fusion Convolutional Neural Network (CNN) and a Gated Recurrent Unit (GRU). Firstly, the mineral, chemical composition and particle size distribution of gold tailings were analyzed, and their leaching toxicity was tested according to relevant standards to ensure their safety and stability as concrete materials. Subsequently, the gold tailings concrete dataset was constructed through experiments and applied to the training and validation of the model. In order to further verify the predictive ability of the model, it was applied to real engineering cases. The results show that the proposed model exhibits high accuracy in both the training and testing process, and is capable of effectively predicting the compressive strength of the gold-tailings-based concrete. The actual engineering cases show that the error range between the predicted and measured compressive strength of concrete with 20%−40% gold tailings is −4.1%−5.7%, which further proves the potential of the model to be applied in engineering practice.

Taking the filter membrane of a microbial sampler as the research background, a porous step model was used to simulate the structure of the microbial sampling membrane. The internal flow field of the sampler was numerically investigated using the finite element method. Simulations were conducted to analyze the filtration pressure drop, flow velocity variation patterns, and the distribution of velocity and pressure contours within the sampler. The influence of fluid inlet velocity parameters on the performance of the deep-sea microbial sampling device was studied, thereby determining the optimal fluid inlet velocity parameters for the deep-sea microbial sampler. The results show that under the conditions of inlet velocity of 0.1, 0.2, 0.3, 0.4, 0.5, the pressure drops on both sides of the filter membrane of the sampler are 36, 82, 139, 207, 288 Pa, and the differences of flow velocity on both sides of the filter membrane of the sampler are 0.05, 0.16, 0.26, 0.34, 0.38, respectively. The pressure drop across the filter membrane of the sampler and the velocity difference between both sides of the membrane increase with the increase in the inlet velocity, while the pressure behind the filter membrane decreases as the velocity increases. When the fluid inlet velocity is 0.3, the filtration performance of the sampler reaches its optimum. Numerical simulation results can provide a theoretical basis for researching the filtration performance of deep-sea microbial sampling devices.

The large range multi-level and multi-stope strength mining in deep metal mines causes frequent and violent adjustment of surrounding rock stress field, which is easy to induce engineering rock mass disasters such as spalling and caving. Aiming at the problems of high cost of multi-point test and poor timeliness in the inversion process of mining-induced stress field, an improved density clustering algorithm based on dynamic optimization of neighborhood search range of microseismic source radius was proposed, which divided microseismic activity into multiple clusters with high internal correlation. On this basis, the natural neighbor interpolation method was adopted to establish the inversion method of "microseismic signal apparent stress – microseismic event apparent stress – microseismic cluster apparent stress – mining-induced stress field". The engineering practice of deep metal mine shows that the average contour coefficient of microseismic data clustering cluster family is 0.56, which verifies the reliability of the improved density clustering algorithm for microseismic data clustering, and the apparent stress concentration area obtained by this method is highly consistent with the actual ground pressure disaster event area. The research results can provide theoretical and technical support for the identification and risk control of potential dangerous areas of mining-induced ground pressure in deep metal mines.

The gently inclined thick ore body below －675 m level in Chengchao Iron Mine is mined by two-step mining with sublevel drilling open stoping and subsequent filling. According to the theoretical analysis results, three numerical simulation research schemes of mine room (pillar) span were formulated. The stress, displacement and plastic zone distribution of each scheme were compared and analyzed, and the influences of different spans of mine room (pillar) on the stability of stope was explored. The results show that when the span of mine room (pillar) is 15 m, the maximum tensile stress of the goaf roof is less than 2 MPa, and the maximum compressive stress is less than 5 MPa, which is within the allowable range of the rock mass. The maximum subsidence value of the goaf roof is less than 5 cm, which does not exceed the allowable limit displacement. There is no large area of plastic zone penetration in the structural plane of the goaf. When the span of the mine room is 18 m, the roof settlement is 52.3 mm (over the warning value of 50 mm), which need to be supported. It is determined that the span of mine room (pillar) is 15 m as the optimal scheme. The maximum vertical subsidence value of the monitoring points in the center of the roof is 0.033 7 m, and the maximum vertical subsidence value of the monitoring points in the center of the surrounding rock is 0.031 4 m. According to the evaluation results of the influence of ground deformation on the buildings (structures), the mining of the ore body below －675 m level will have a slight impact on the surface buildings (structures), and the influence degree is within the allowable range, which ensures the safety of the new Chengchao Bay country on the surface of Chengchao Iron Mine.

In order to study the degradation law of mining height on the strength of overlying rock, taking the determination of reasonable initial mining height in the initial mining stage of Guojiawan 51208 working face as the background, the fracture density function was used to measure the degradation of mining height on the strength of overlying rock, and the theoretical model and numerical analysis model of the strength degradation of overlying rock under different mining heights were established. Combined with the parameters of the working face, the influence law of mining height on the stability of overlying rock was obtained. The results show that with the increase of mining height, the rotation angle gradually increases, the thickness of the rock layer loaded by the bearing layer gradually decreases, and the rock layer is more prone to rotary deformation and instability. When the advancing length of the working face is 40 m, the maximum displacement of the overlying rock and the plastic zone distribution of the basic roof are basically the same for the mining height of 2.5 m and 4.0 m. When the advancing length of the working face is 50 m and above, the maximum displacement of the overlying rock with a mining height of 4.0 m is more than10%higher than that with a mining height of 2.5 m, and the plastic zone of the basic roof is obviously increased. The peak value of advance support pressure increases with the increase of mining height, and the first weighting step decreases with the increase of mining height. The initial weighting step of 2.5 m mining height is 70 m. The initial mining height is 2.5 m, and the mining height is gradually adjusted to the actual coal thickness（4.0 m）after the first weighting of the main roof. The initial mining of the working face is more efficient and safe. The research results can provide reference for the initial mining roof control problem of the working face.