Through the true triaxial loading-disturbance tests of granite with different structural planes, the acoustic emission system was used to monitor the rock failure process, and the failure characteristics and acoustic emission evolution characteristics of rocks with different structural planes were explored. The relationship between structural surface parameters and rock sample failure outcomes was analyzed, and macroscopic and microscopic differences of the three failure results were compared. The results are concluded as follows. (1) When the length of the structural plane is 20 mm and 30 mm, the rock sample is prone to crack-induced failure. When the length of the structural plane is 50 mm, the rock sample is prone to rockburst. With the increase of the length and dip angle of the structural plane, the failure of the rock sample tends to change from shear failure to internal shear and external splitting. When the length of the structural plane is short, the dip angle has a weak influence on the failure mode of the rock sample. Structural plane grouting will weaken the intensity of rock failure. (2) The proportion of tensile and shear areas in RA-AF density diagram of rock samples with three kinds of failure results is different. (3) With different structural plane parameters, the crack distribution and amplitude exhibit significant differences in the upper, middle, and lower positions of the rock sample's structural plane. (4) Based on the analysis of the variation characteristics of the coefficient of variation CV(r) of b value and r value, it is found that there are significant differences between the formation of shear cracks and tensile cracks in granite with different structural plane parameters under dynamic disturbance. This study can provide theoretical guidance for the prevention and control of mining-induced disasters in deep rock mass with structural planes.

To reveal the influences of bedding angle and chemical corrosion under acidic environment on the mechanical behavior of carbonaceous slate, triaxial compression tests under chemical acidic environment corrosion were performed on bedding carbonaceous slate specimens collected from the side walls of large-deformation roadways. A continuous-discontinuous coupled numerical model based on the combined discrete element method (DEM) was developed, and numerical analysis under extended working conditions was conducted using the improved flexible membrane radial loading method, with results validated against physical test data. The results are concluded as follows. Under triaxial compression, the rock specimens develop inclined and penetrating fractures, accompanied by shear slip failure along the direction of the bedding angle. The corrosion of the chemical solution accelerates interlayer slip, leading to the gradual emergence and concentration of fractures. The peak strength of carbonaceous slate firstly decreases and then increases with the increase of bedding angle, reaching the minimum value especially when the bedding angle ranges from 15° to 45°. This phenomenon is mainly attributed to the wing crack propagation caused by shear slip and local stress concentration at this angle range. The chemical acidic environment significantly accelerates the structural deterioration of the slate. After 90 days of corrosion, the peak strength of the samples decreases by 36％ to 44％, and there is a negative correlation between the peak strength and the duration of acidic chemical corrosion. The research results provide a theoretical basis for the stability evaluation and support design of deep roadways.

To explore the status of safe carbon reduction in China's mining industry, the data envelopment analysis (DEA) method was used to measure the safe carbon reduction efficiency of mining industry in 30 provincial-level administrative regions of China from 2014 to 2022. Additionally, social network analysis (SNA) method was utilized to examine the spatial correlation network characteristics in detail. The results reveal that the overall efficiency of safe carbon reduction in China's mining industry exhibits a growth trend with relatively small regional disparities, indicating the green transformation of the mining industry is moving towards a balanced and coordinated safe carbon reduction direction. In terms of the spatial correlation network, while there are some interactions among provincial-level administrative regions, the overall connectivity is weak, and interaction frequency needs improvement. Further analysis shows that different regions play distinct roles in spatial spillover effects. Certain provinces occupy key positions in the network, demonstrating high centrality and significant roles as structural holes and intermediary holes. Moreover, while intra-regional connections are relatively strong, the collaborative effects between regions still have room for enhancement. The research results provide a novel perspective for comprehensively understanding the status of safe carbon reduction in China's mining industry and offers scientific guidance for optimizing policy design and promoting regional collaborative development.

In view of the contradiction between the agility of technical iteration and the lag of system renewal in the emergency system during the intelligent transformation of coal mines, an effective coordination mechanism was constructed to improve the efficiency of safety governance. Specifically, the “technology-system collaborative threshold model” was proposed, and on this basis, a multi-source data fusion framework and a “technology-system-insurance” ternary collaborative mechanism were constructed as follows. At the technology level, the federal learning was used to dynamically optimize the data weight and the OPC UA protocol unified semantics, which effectively solved the heterogeneous problem of downhole sensor protocol. At the system level, the dynamic adaptation rules of AI authority and responsibility were designed. At the insurance level, the linkage mechanism of algorithm filing and liability insurance was pioneered to systematically fill the gap of intelligent emergency system. Empirical studies demonstrate that at Yujialiang Mine, multi-source data fusion can increase operational efficiency by 80％ and reduce the false alarm rate for gas emissions to 4.3％. Meanwhile, at Hongqinghe Mine, introducing an optimization mechanism with an environmental priority coefficient, can achieve a 60％ reduction in ventilation response time. In summary, the constructed coordination threshold model and its implementation mechanisms provide a crucial quantitative tool for transforming coal mine safety governance from a passive-reactive approach to a proactive-defensive one. Furthermore, it establishes an innovative paradigm of “technology-institution-ecology” synergistic development, thereby advancing the construction of green mines.

With the increase in the mining depth of mines with high in-situ stress, the problem of low intelligence level of traditional drilling rigs has become a challenge for borehole pressure relief. To this end, an innovative design scheme for an intelligent mine pressure relief drilling rig was proposed. This intelligent drilling rig integrates four key components, including the chassis walking unit, the drilling execution unit, the support lifting unit, and the drill rod magazine unit. Among them, the unique automatic drill rod changing device realizes the flexible displacement of the drill pipe in the precision track system through the cooperation of the horizontal and the vertical pushing mechanism. With the cooperation of the fixing groove and the clamping mechanism, the automatic drill rod changing operation process can be completed efficiently and accurately, significantly improving the automation level and continuity of the operation. At the same time, the advanced in-drill information monitoring and analysis system built into the drilling rig can capture and analyze multi-source data during the drilling process in real time, laying a solid foundation for precisely adjusting the drilling parameters and promptly warning of potential risks. The overall structure of the intelligent drilling rig, as well as the structure of each unit were introduced, subsequently, based on the analysis of a large amount of drilling information, the excellent performance of the intelligent pressure relief drilling rig was comprehensively verified. Compared to the traditional drilling rigs, the operational safety of intelligent drilling rigs has been improved, equipment failure rates have been reduced, the drilling time and auxiliary operation time have been shortened, and its drilling efficiency has been improved.

Underground mine drill bit detection is a critical link to ensure the safety of underground mine operations, but existing deep learning-based detection models have shortcomings in feature extraction capability and detection efficiency. To improve detection performance, a drill bit detection model named YOLOv8-WGB based on the improved YOLOv8n was proposed. Firstly, the traditional CIoU loss function was replaced with the WIoU loss function, which enhanced the focus on key features. Secondly, the global attention mechanism (GAM) was introduced, which enhanced the global information interaction capability of the YOLOv8 backbone network, reduced information loss, and amplified global dimensional features. Finally, the BiFPN structure was added, which assigned different weights to feature maps of different scales when fusing them, so as to achieve optimal feature fusion. The test results show that YOLOv8-WGB improves precision, recall, and mAP@0.5 by 7.1%, 11.1%, and 8.3% respectively compared with YOLOv8n, and its inference speed is 173 frames/s faster than that of YOLOv8s, exhibiting higher detection speed. On the basis of maintaining the model complexity and inference speed advantages of YOLOv8n, the proposed model has significantly superior detection accuracy to YOLOv8s. It effectively balances detection accuracy and speed, and exhibits excellent performance in the underground mine drill bit detection task.

In the field of UAV-based open-pit mine surveying, most previous studies adopted only one type of UAV for data collection and analysis, without considering the issue of internal and external operation efficiency. Therefor, taking Xiangxiang Dolomite Mine as the engineering background, two types of UAVs, namely Matrice 300 RTK and Mavic 3E, were adopted. External data collection with four different flight heights was carried out in the mining area. In the internal data processing, four image control point layout schemes were selected respectively under each working condition to analyze the measuring point accuracy under different flight heights and different numbers of image control points. Meanwhile, statistical analysis of the internal and external time-consuming under each working condition was conducted, and the surveying operation efficiency of UAV-based open-pit mine was analyzed. The results show that the overall accuracy of the M300 RTK is higher than that of the Mavic 3E under the same working conditions, and as the flight heights increase, the accuracy continuously decreases. Under the working conditions of the same flight height, as the number of image control points increases, the accuracy of the two types of UAVs continues to improve, but the improvement range is not significant. Under the working conditions of the same flight height, the external time-consuming of the two types of UAVs is basically consistent, but the internal time-consuming of the Mavic 3E is only 3.8％ to 8.8％ of that of the M300 RTK. In non-coal open-pit mines with large height difference, the adoption of the Mavic 3E's real-time terrain-following flight combined with the image-control-free mode not only eliminates the field operation step of laying out image control points but also significantly reduces internal time-consuming. While ensuring accuracy, it can greatly improve the surveying operation efficiency of UAVs in mines.

The scarcity of samples seriously restricts the promotion and application of deep learning technology in the field of intelligent recognition of rock and mineral. Therefore, an intelligent incremental method for image samples of rock and mineral based on an improved deep convolutional generative adversarial network (DCGAN) was proposed. By introducing self-attention mechanism to enhance the details and authenticity of generated images, using loss function with gradient penalty mechanism (WGAN-GP) to improve training stability, expanding network structure to enhance model performance, and combining various data augmentation techniques to expand sample diversity and quality, the problem of sample scarcity has been effectively alleviated. The experimental results show that the proposed method outperforms comparative models in evaluation indicators such as peak signal-to-noise ratio, Fréchet initial distance, and structural similarity index. The loss function curves show that the model training process is stable and has a good convergence performance. The clustering analysis results indicate that images generated by the model are superior to the original images in terms of complexity and diversity. The survey results show that the model can generate high-quality and highly realistic images of rock and mineral. The research effectively alleviated the problem of scarcity of rock and mineral image samples, laying the foundation for the intelligent development and practical application of rock and mineral image generation.

A large number of iron tailings can not be fully and effectively utilized due to their low content of useful minerals, weak magnetism, fine crystal particle size and complex and difficult separation. Therefore, in order to solve this problem, a comprehensive utilization study was conducted on low-grade ore tailings from a beneficiation plant in the central Shandong region. Firstly, the properties of the original ore were analyzed, tested, and screened for analysis, and it was found that the main iron containing mineral in the original ore is magnetite. Based on the properties and characteristics of the ore, three process schemes were tested on the raw ore, namely screening－gravity concentration－medium magnetic separation－grinding－medium magnetic separation－gravity concentration process scheme (Scheme Ⅰ), screening－gravity concentration－medium magnetic separation－grinding－medium magnetic separation－flotation concentration process scheme (Scheme Ⅱ) and screening－gravity concentration－medium magnetic separation－grinding－medium magnetic separation－gravity concentration－gravity concentration－gravity tailings flotation concentration process scheme (Scheme Ⅲ). The final results show that Scheme Ⅱ has the best sorting effect, achieving a concentrate with yield of 20.46％, iron grade of 50.93％, and iron recovery rate of 44.10％. Among these ores, there are iron minerals embedded with fine particle size and closely coexisting with silicate iron minerals with relatively strong magnetism. To obtain higher grade iron concentrates, gravity separation or flotation shows obvious sorting advantages compared to magnetic separation. Therefore, Scheme Ⅱ or Scheme Ⅲ can be selected according to the actual situation.

In order to give full play to the alkali-activated reactivity of industrial solid waste raw materials, determine the degree of reaction and the type of gel formed, and optimize the mechanical properties and durability of alkali-activated cementitious materials, the effects of Ca, Fe, S, Cl, Mg and other elements on the mechanical properties, corrosion resistance and microstructure of industrial solid waste-based alkali-activated cementitious materials were summarized. The results show that the synergistic effect of Ca and S in the AACMs system complicates the reaction process. Calcium-free and low-calcium AACMs have good acid resistance, while calcium-rich AACMs are prone to decalcification under acid attack, which weakens the structural strength. Compared with the materials rich in silicon, chloride and calcium, the higher iron content has an adverse effect on the compressive strength and corrosion resistance of AACMs. The resistance to chloride ion penetration of ACCMs is related to the chloride ion binding capacity of its skeleton, the presence or absence of chlorine in the system and the pore structure. At the same time, alkaline cations also have a significant effect on the chloride ion binding capacity and resistance to chloride ion penetration of ACCMs. Finally, the problems and future development directions of alkali-activated cementitious materials were discussed, which can provide reference for further research on industrial solid waste-based alkali-activated cementitious materials.