According to the drainage data of 7 horizontal wells in Tunlan block, the productivity characteristics of horizontal wells were analyzed. It is found that there are great differences in water production and productivity of horizontal wells. The influence of geological factors on productivity was analyzed from the aspects of coal seam buried depth, coal seam thickness and roof and floor lithology. The influence of engineering factors on productivity was analyzed from the aspects of fracturing curve characteristics, average displacement, fracturing fluid amount and sand addition. The influence of drainage factors on productivity was analyzed from the aspects of wellhead casing pressure, drainage continuity and water production. The results show that the gas production of horizontal wells is affected by the water production. In general, there is a positive correlation between the two, but too much water production will lead to a decrease in gas production. Hydrogeological conditions have an impact on the productivity of some horizontal wells. The increase of coal seam burial depth, coal seam thickness and roof mudstone thickness is conducive to the extraction of coal bed methane. The gas production of horizontal wells with low displacement fracturing is low, the fracturing fluid amount and sand addition should also be kept in a certain range, and excessive use will lead to a decrease in gas production. The descending and stable fracturing curves account for 81.4% of the horizontal well fracturing curves. The higher the proportion of the curve, the greater the gas production. In the process of drainage, the long maintenance time of the stuck pump, the decrease of the wellhead casing pressure and the high water production will reduce the gas production. The grey correlation analysis shows that the gas content of coal seam is the main controlling factor of horizontal well productivity. The research results can provide reference for the development of horizontal wells in Tunlan block.

To address the problem of pore wall instability in the fractured formation during geological drilling at Pingpo Coal Mine in Gansu Province, the mineral compositions, microstructures and pore characteristics of mud shale formation in coal measure strata were comprehensively analyzed. It is found that the clay minerals are mainly composed of illite and montmorillonite with high contents, and the degree of mineral particle cementation is low. At the same time, the large pores and medium pores account for more than 80% of the total pores volume, which is easy to cause pore wall collapse under the erosion of fluid medium. Nano materials and micro-foam agents with excellent performances were selected through laboratory experiments, and a kind of nano silicon-based micro-foam drilling fluid suitable for fractured formation drilling was developed based on the low solid content drilling fluid. The results show that the drilling fluid has a good fluidity and filtration, which can effectively suppress the deformation of coal measure strata and the collapse of pore walls, and has strong infiltration sealing performance and inhibition capabilities. Through the field application of the new drilling fluid in coal measure strata drilling, the pure drilling rate of the drilling hole was increased by 40%, the core recovery rate reached 95% and the monthly efficiency was increased by 72%, which proved that the new drilling fluid has a good wall protection and anti-collapse function in the fractured formation of coal measure strata.

The visualization of rockburst risk is of great significance to the exploitation of deep phosphate rock resources. Firstly, through the uniaxial loading test of aluminum metal, it is proved that the temperature change caused by thermo-mechanical coupling can be detected by MicroIIIS series 640 infrared thermal imager, and the increase of stress is the core reason for the increase of thermo-mechanical coupling temperature. Then, aiming at the problem that the infrared image in the actual phosphorus mine roadway is blurred and the irrelevant high temperature body intervenes to affect the effective infrared radiation characteristics information in the infrared image, the original gray image is enhanced by the adaptive gray scale stretching algorithm, so that the information entropy, histogram equalization and global contrast are optimized. The nonlinear mapping pseudo-color transformation based on temperature width adjustment is applied to enhance the visual effect of rockburst risk visual detection in actual phosphate rock roadway. Finally, through the field test of phosphate rock roadway, it is proved that the DMYB-03 handheld rockburst detector after infrared image enhancement can not only show the details of phosphate rock roadway and the infrared radiation characteristics caused by stress concentration more clearly, but also divide the stress concentration degree into stable, medium stress concentration and high stress concentration through the color brightness index of risk area and safety area. The information entropy of high stress concentration is 26.55% higher than that of stable image, the average brightness is 25.08% higher, the brightness histogram equalization is 8.36% higher, and the global contrast is 243.38% higher, semi-quantitatively reflects the degree of stress concentration of rock mass. The research results provide technical support for the visual detection of rockburst risk in mining, transportation and other fields.

On the basis of revealing the formation mechanism and development mechanism of the slope in the latent sliding zone of Shuguang Gold-Copper Mine, the treatment work of slope in the latent sliding zone was carried out. The formation factors and processes of the latent sliding zone were studied through on-site engineering geological investigation and limit equilibrium analysis. Based on engineering geological investigation, the formation factors of the latent sliding zone were divided into internal and external factors, and the formation process was divided into three stages, including the generation, development, and penetration of slope cracks. Combining the treatment conditions, a treatment scheme with large-diameter anti-slip piles has been formulated. Through engineering geological investigation and parameter back calculation, the rock mass parameters and the position of sliding surface of the latent sliding zone have been determined. Based on the residual sliding force calculation, a single row of circular anti-slip piles with a diameter of 800 mm, a length of 32 m, and a hole spacing of 4 m will be constructed on the 590 m platform. Three limit equilibrium analysis methods were used to verify the stability of the post-treatment slope, and the results show that anti-slip piles can effectively resist sliding. The research results can provide a reference for safety production and slope treatment in mines.

In order to find out the influence of water disaster in 2-3 coal seam roof mining in Gaotouyao Coal Mine, the “three-graph and double-prediction method” coupled with GIS and AHP was used to construct the water abundance evaluation model of 2-3 coal seam roof water-filled aquifer. Comprehensive comparison of FLAC3D numerical simulation, empirical formula calculation, analogy measurement and other methods, the traditional calculation formula of fractured water-conducting zone height was modified. On this basis, the classification and zoning map of water gushing (inrush) risk of 2-3 coal seam roof water-filled aquifer was obtained by compound superposition. Numerical simulation was used to predict the water inflow of Gaotouyao Coal Mine. The results show that the risk of water inrush in the central and southwest of Gaotouyao Coal Mine is higher, and the risk of water inrush in the northwest and east is lower. The normal water inflow of Gaotouyao Coal Mine is 465 m³/h, and the maximum water inflow is 558 m³/h. In view of the water disaster problems faced by the mine in the future, the feasible prevention and control measures have been put forward comprehensively.

In response to the randomness and fuzziness of water inrush risk in coastal gold mines, as well as the complex correlation among evaluation indicators, a GRA-VIKOR evaluation model of water inrush risk in coastal gold mines based on moment estimation theory was established to make the risk assessment more scientific. Firstly, 10 typical influencing factors were selected from two aspects of engineering geology and hydrogeology, and an evaluation indicator system was constructed. Secondly, four single weighting methods, namely IAHP method, G1 method, EM method, and I-CRITIC method, were used to calculate the subjective and objective weights of the indicators. Based on moment estimation theory, the integrated optimization method of subjective and objective weights was improved. According to the GRA-VIKOR principle, the risk status of water inrush was analyzed and determined by calculating the comprehensive evaluation values. Finally, the model was applied to the risk assessment of water inrush in 9 middle sections of a gold mine in Shandong Province, the risk level and ranking of water inrush in each middle section were determined. The results indicate that the evaluation results of the model are accurate and reasonable, in line with reality, and can provide reliable ideas for the risk evaluation of water inrush in coastal gold mines.

In order to improve the prediction accuracy of multi-source index risk identification in goaf, based on the particle swarm optimization (PSO) algorithm to optimize the input weights and thresholds of the extreme learning machine (ELM), a PSO-ELM method was proposed to identify the risk of multi-source indicators in the goaf. Nine indicators, including mining depth, mining method, goaf height, the maximum exposed area, the maximum exposed height, the maximum exposed span, mining pillars, measured volume, and control rate, which affect the risk of goaf in a mine in South China, were used as the main input values, and 135 measured goaf data were used as an example to predict the hazard level of goaf. The research results show that the prediction accuracy of radial basis function and traditional extreme learning machine is 88% and 92%, respectively, while the prediction accuracy of the extreme learning machine model optimized by particle swarm optimization algorithm is 96%. Compared with radial basis functions and traditional extreme learning machines, the PSO-ELM algorithm performs better on the test sets, has a better stability and generalization ability, possess good guiding significance and practicality, and is suitable for practical engineering applications.

In extensive cold region engineering, freeze-thaw failure of rock masses is a natural disaster that can’t be ignored. In order to explore the influence of freeze-thaw environment on the engineering stability of rock masses in cold regions, uniaxial compression tests were conducted on red sandstone in cold regions under different freeze-thaw cycles (0, 10, 20, 30 times) and saturation levels (0, 20%, 40%, 60%, 80%, 100%), and SEM microstructure analysis was performed on the rock specimens. The results show that with the increase of freeze-thaw cycles and saturation, the crack closure stress σ_cc, initiation stress σ_ci, damage stress σ_cd, peak stress σ_c, and frost resistance coefficient K_fm of red sandstone gradually decrease, and internal damage gradually intensifies. The failure mode of red sandstone specimens gradually changes from oblique shear failure and tensile failure to X-shaped failure, accompanied by a small amount of fragment peeled off. Recurrent freeze-thaw cycles leads to the expansion of original microcrack and the emergence of new cracks inside the rock, revealing the damage evolution characteristics of sandstone under different freeze-thaw cycles. The research results can provide a certain reference for the study of rock engineering stability and rock mechanics properties in cold regions.

To explore the damage degradation and landslide instability mechanism of weathered red-bed soft rock under the action of acid rain, the stress-seepage coupling creep test of red-bed soft rock under repeated erosion of acid solution was carried out, and the mechanical behavior and damage evolution mechanism of rock coupling creep were analyzed. By analyzing the time-strain curve of rock creep under different test conditions and the relationship between rock creep rate and permeability before and after hydrochemical damage, the mechanism of rock damage degradation under the coupling of chemical-stress-seepage and the main factors affecting rock damage were explored. The results show that the hydrochemical action will reduce the triaxial strength of the rock, with a minimum reduction of 36.31% and a maximum reduction of 48.03%. At the same time, the rock creep curve after repeated erosion by hydrochemical action is more divergent. There is a good positive correlation between rock damage law and stress level under different hydraulic conditions. The degree of hydrochemical damage is much higher than that of seepage and stress. After hydrochemical damage, rock seepage and stress will aggravate rock damage degradation. Under the action of high seepage pressure, the creep effect of rock is more obvious. Finally, the chemical damage factor is introduced to establish a triaxial nonlinear damage creep model under the stress-seepage coupling effect including chemical damage, which can describe the fluid-solid coupling creep characteristics of red-bed soft rock under the action of hydrochemical action, and the fitting effect is good.

In geotechnical engineering, instability failure often occurs in multi-layer rock mass at the same time, and its different combinations, lithology and other factors will affect the development trend of cracks. The discrete element PFC2D software was used to establish the sample model, and the uniaxial compression experiment was simulated to reveal the influence of the proportion of soft and hard rock strata on the mechanical properties, failure characteristics and energy evolution law of rock samples. The results are drawn as follows. Firstly, the proportion of hard rock in the layered composite rock sample is the main factor affecting the peak strength of the rock sample. As the proportion of hard rock increases, the peak stress of the rock sample gradually increases. Secondly, the number of microcracks and the total input strain energy of the layered composite rock sample are the minimum when the thickness ratio (R) of the adjacent rock layer is the smallest, indicating that the decrease of R will lead to the early failure of the rock sample. Thirdly, the elastic energy consumption ratio K is introduced into the energy analysis of layered composite rock samples, and it is verified that the increase of K value would be used as an early warning signal for rock deformation and failure. Fourthly, the main failure mode of the layered composite rock sample is tension-shear mixed failure, and the macroscopic cracks generally originate in soft rock part. Then, due to the sharp increase of the transverse strain of the soft rock part, the cracks penetrate into the hard rock part, revealing that the failure of the rock sample is a process of gradual loss of bearing capacity. The research results can provide a reference for practical engineering.