There are well-developed faults in an open-pit molybdenum mine, which may lead to slope instability during open-pit mining. Three typical faults running through the ultimate boundary were selected to investigate progressive instability mechanism of the slopes in the open-pit mine under coupling effect of multiple faults by integrating Rhino+Griddle 3D geological modeling and FLAC^3D multi-field coupling simulation. Furthermore, the safety factors of slopes under different working conditions were calculated, and relatively unstable slopes were also identified for addressing the subsequent safety of the ultimate boundary. The simulation results show that, plastic failure and tensile stress concentration will occur in the slope at the junction of Zone C and Zone D, while local plastic failure will occur in the faults, which, however, won't induce instability failure in the surrounding slopes; there won't be much displacement in the slope at the ultimate boundary, with the maximum displacement (2.44 cm) at the toe of slope; the safety factors of the ultimate boundary of open pit under natural working conditions and rainfall are 2.48 and 2.36 respectively, indicating that the slope is stable. However, slope at Zone E with F1 fault is considered to be with a relatively instability, for which comprehensive measures including real-time monitoring, prevention and reinforcement can be taken to ensure the safety.

Based on the causes of historical dam failure accidents in tailings ponds, a safety grade evaluation index system was built for tailings ponds. The correlation between safety evaluation indexes was demonstrated with Pearson correlation analysis method. An in-depth analysis was conducted for the distribution of sample data to confirm that the sample data met the requirements of the Kriging method for data distribution. Then, the weight of each safety evaluation index was calculated with entropy weight method. The two safety evaluation indexes with higher sensitivity were selected as the X-axis and Y-axis of the Kriging geostatistical model. A geostatistical model was established by using the Kriging method and was verified with empirical case data. The results show that the predicted safety grades of 12 groups of tailings pond data with known safety grades in the geostatistical model are completely consistent with the actual safety grades, which verifies the feasibility and high accuracy of this method.

To analyze the damage effect of drilling and blasting in construction on the surrounding rock of roadway, a Hopkinson pressure bar (SHPB) test and numerical simulation were carried out to investigate the dynamic response characteristics of saturated mudstone under one-dimensional impact load. Under impact with different strain rate, rock samples undergo four deformation stages, including compaction, quasi-elasticity, yielding and unloading. The saturated mudstone has its dynamic elastic modulus in a linear relationship with the strain rate. As the strain rate increases, dynamic fragmentation of the rock samples is significantly enhanced, and the energy dissipation of saturated mudstone also correspondingly increases. The mean error between the peak stress obtained from the numerical simulation and the experiment is 7.9%, which accurately reflects the dynamic characteristics, such as morphology of mudstone fragmentation and dynamic stress-strain relationship, under impact with different strain rates. The strain rate effect of mudstone indicates that blasting disturbance will continuously change its energy state, so dynamic monitoring should be strengthened for the deformation of surrounding rock during tunneling.

Based on the mining environment and stress characteristics of backfill in the upward horizontal slicing and filling method, the influence of wetting-drying (WD) cycles on the mechanical properties of cemented tailings backfill (CTB) was explored. An ultrasonic test and a mechanical property test were conducted for CTB samples with different cement-tailings (c/t) ratios after different WD cycles to investigate variation in the velocity of longitudinal waves, compressive strength, and failure mode. The results show that after WD cycles, the CTB samples with c/t ratios of 1∶4, 1∶6 and 1∶8 all had a decreased ultrasonic wave velocity, and as the number of WD cycle increases, CTB samples with different c/t ratios had enhanced plastic deformation capacity but decreased strength. It is shown that WD cycle has a small impact on backfill with a higher c/t ratio. And the CTB without WD cycle mainly experiences tension-shear failure, while CTB after WD cycles has multiple axially parallel cracks penetrating vertically, which significantly increase after more WD cycles. It is found that the c/t ratio is a key factor for CTB to resist deterioration caused by WD cycles. In mining operation, the c/t ratio for different filling areas can be reasonably optimized to reduce filling costs.

In order to explore the stability and durable service of a blend of construction waste and red clay under load, a California Bearing Ratio (CBR) test, rebound modulus test and failure strength test were performed to determine the appropriate content of red clay in the blend. The influence of compaction degree, confining pressure, stress ratio and loading times on the macroscopic permanent deformation of the blend was comprehensively analyzed by triaxial tests, and the evolution of internal structure during the deformation process of the blend was simulated by discrete element method. The results show that with red clay at an appropriate content of 65%, the blend of construction waste and red clay can have its permanent deformation increase gradually as the stress level becomes higher, and decrease gradually with the increase of confining pressure and compaction degree. During the deformation process, the internal shear stress, coordination number and slip rate of the blend all increase gradually as the stress level becomes higher. Based on the test results and in consideration of stress state, including confining pressure, failure strength and loading stress, physical state, such as compaction degree, and loading times, a permanent deformation prediction model was proposed and validated. With a correlation coefficient of 0.89, the model is considered to have a “relatively good fit”.

As for fracture development and potential instability in the open-pit bench slopes in Dabaoshan in Guangdong Province, the blockiness was proposed to be taken as a quantitative measure of rock mass integrity. Based on the existing data, 72 potentially unstable zones on the western slope were identified. The structural characteristics of the rock mass were analyzed based on investigation of fracture occurrence characteristics and 3D fracture network modeling, and rock mass integrity was also evaluated by blockiness analysis. It is found that there are fully-developed fractures with good ductility in all the zones, creating the structural conditions for block formation; in 84.9% of the zones, the inclination angle of superior fractures and the slope inclination angle are supplementary, which can significantly reduce sliding risks; in 97.3% of the zones, the blockiness of rock mass is less than 27, indicating mild blockiness, good integrity of rock mass, and overall high stability. There are only rock mass with blockiness over 27 in Zone 6^# and Zone 46^#, presenting moderate blockiness. In these two zones, the rock mass structure is significantly degraded, leading to reduction in damage tolerance. It is concluded that there are potential risk of slope instability in those two zones, which should be taken as the focus in the subsequent disaster prevention and control for the slopes in Dabaoshan.

Homogenization of backfill slurry cannot be automatically evaluated during stirring process. To solve this problem, the Yolo-v7 algorithm was improved by using the CBAM and the SPD-Conv, and the improved Yolo-v7 algorithm was then employed to automatically evaluate the homogenization of backfill slurry. The research results show that the improved Yolo-v7 algorithm demonstrates a significantly improved performance, with the accuracy, recall rate and mean average precision (mAP) all increased by 17.5, 28.8, and 32.4 percentage points respectively. Analysis of slurry parameter sensitivity indicates that slurry concentration is the principal factor affecting evaluation of slurry inhomogeneity, followed by the cement-tailings ratio. Moreover, the homogeneity of backfill slurry with high concentration can be significantly improved by prolonging stirring time.

To reveal the relationship among uniaxial compressive strength, tensile strength and shear strength, 159 sets of rock mechanics data collected from 38 mines were taken as samples. Firstly, the relationship between uniaxial compressive strength and tensile strength was characterized by data analysis, and the range of uniaxial compressive and tensile strength in the formulae for calculating shear strength was calibrated. Then, a BP neural network prediction model was constructed, with which the shear strength of rocks can be predicted with uniaxial compressive strength and tensile strength as the input layer. The results show that: the uniaxial compressive strength of rock is 7.6-27.8 times tensile strength; the cohesion value obtained from multiple comparison tests is 1.0-2.2 times that from the original empirical formula, and the tangent value of internal friction angle is 0.35-1.00 times that from the original empirical formula, presenting high confident and reliable results; based on comparison between the predicted value with BP neural network and the actual value, the cohesion deviation ranges from -10.53 MPa to 20.12 MPa, and the internal friction angle deviation ranges from -18.43° to 28.97°, exhibiting an overall normal distribution.

To enhance the positioning accuracy of seismic source in underground mining and solve the problem of spatiotemporal changes in the wave velocity field caused by excavation disturbance and complex structures, seismic source relocation with real-time inversion of wave velocity field as a prior was proposed, which combined ray tracing and quasi-Newton iteration algorithm to realize dynamic update of wave velocity field and high-precision positioning of seismic source. The validity of this method was verified by synthetic monitoring and field testing. The synthetic monitoring results showed that compared to the least squares procedure, the real-time inversion resulted in the average positioning accuracy enhanced by 49.8% and the accuracy of wave velocity inversion exceeding 95%. In the field testing, a 180 m×180 m to-be-filled area in mining operation was taken for imaging target. Compared to the least squares procedure, this method brought the average error in positioning reduced by 7.074 m, and the accuracy of wave velocity inversion in each region exceeding 95%. The research indicates that this method not only is suitable for seismic source positioning in microseismic monitoring, but also can be used as a geophysical method for detecting passive imaging of goaf areas.

The polymetallic ore from Shizhuyuan of Hunan Province was taken for research. Based on the characteristics of uneven coarse-fine dissemination of minerals and insufficient deliberation of Mo-Bi sulfide ore for flotation, an enhanced recovery technology by stepwise separation was developed for the low-grade Mo-Bi sulfide ore. The results show that an experiment by adopting such stepwise separation technique produced a bulk flotation concentrate grading 1.31% Mo and 2.83% Bi at corresponding recoveries of 88.43% and 85.99%, presenting 5.34 and 4.80 percentage points up, respectively, compared to on-site conventional bulk flotation process. This creates favorable conditions for the following tungsten flotation from the bulk flotation tailings. It is shown that a closed-circuit tungsten flotation can produce a tungsten concentrate grading 44.64% WO₃ at 84.12% recovery.