As the foundation for regional landslide risk assessment, landslide susceptibility prediction (LSP) is a prominent and challenging topic in global landslide disaster prevention and control research. This paper systematically reviews critical issues in LSP, including model selection, identification and integration of conditioning factors, determination and classification of prediction units, methods for selecting non-landslide samples, optimization of the landslide-to-non-landslide sample ratio, assignment of label values to non-landslide samples, and evaluation methodologies for landslide susceptibility results. The literature review indicates that tree-based models, such as Decision Trees and Random Forests, demonstrate superior performance. The selection and integration of conditioning factors should adhere to principles of comprehensive typology and clear physical significance. Prediction units can be defined through multi-scale segmentation of slope units. Non-landslide samples should preferably be randomly selected from areas characterized by very low or low susceptibility. The optimal ratio of landslide to non-landslide samples can be established through experiments under various conditions. Specific small probability values should be assigned to non-landslide samples. Evaluation of LSP results necessitates a comprehensive consideration of multiple metrics, including ROC accuracy, prediction rate accuracy, and the mean and standard deviation of the susceptibility index. Furthermore, to enhance the accuracy of LSP and validate the cross-regional engineering applicability of the semi-supervised imbalanced theory, this study pioneers the application of the Random Forest model based on this theory in Anyuan-Xunwu County, Jiangxi Province. Experiments were conducted under various scenarios, involving different ratios of landslide to non-landslide samples (ranging from 1:1 to 1:260) and varying label assignments for non-landslide samples. Results indicate that as the ratio increases from 1:1 to 1:180, both ROC accuracy and prediction rate accuracy improve gradually from 0.905 and 0.898 to 0.957 and 0.937, respectively. However, no significant improvement is observed in either metric once the ratio exceeds 1:180. Consequently, the optimal ratio of landslide to non-landslide samples is established at 1:180. Additionally, this study collects data from six landslide events that occurred between 2022 and 2024 to validate the results obtained from the 1:180 ratio. The validation reveals that all six landslides are situated in areas of very high or high susceptibility. This not only confirms the effectiveness and engineering applicability of the semi-supervised imbalanced theory in the mountainous and hilly regions of southern Jiangxi but also provides new insights and technical support for the precise prevention and control of landslide risks.

 In deep coal seams where intact coal interbeds with tectonic coal, the advancement of excavation working faces towards tectonic coal zones, under in-situ stress and gas pressure, is highly susceptible to induced coal and gas outbursts. To investigate the influence of distance on outburst mechanisms during the approach of excavation faces to tectonic coal zones, this study focuses on an outburst-prone coal seam from a mine in Henan. A “tectonic coal mass” experimental model was established to simulate tectonic coal zones embedded within intact coal. Using a self-developed true triaxial coal and gas outburst simulation test system, outburst simulation experiments were conducted at varying approaching distances (L) ranging from 0 to 35 mm between the excavation face and the tectonic coal zone. Throughout the experiments, acoustic emission (AE) technology and temperature sensors were employed to monitor the entire outburst process. New parameters, including the cumulative energy growth rate at each stage and the primary and secondary peaks of AE energy, were defined. Based on the relationship between the primary and secondary AE energy peaks, precursor coefficients for low, medium, and high outburst risks were proposed. The study analyzed AE evolution patterns and temperature characteristics within the tectonic coal zone at different approaching distances. The research findings indicate: (1) A critical approaching distance exists when the excavation face approaches tectonic coal zones. Under consistent in-situ stress conditions, the unit outburst intensity demonstrates a quadratic relationship with the approaching distance, initially increasing and then decreasing. A critical value of unit outburst intensity is observed at an approaching distance of 20 mm. (2) During the outburst incubation stage, the cumulative energy growth rates (K?, K?, K?) exhibit a distinct two-phase trend in relation to the approaching distance: a positive correlation between 10–20 mm and a negative correlation between 20–35 mm. The energy release intensity (K?/K?) during the late triggering phase is 7.17–15.59 times greater than that during the early phase (K?/K?). The approaching distance and AE energy release initially increase before subsequently decreasing. (3) The temperature increment during outbursts in the tectonic coal zone first rises and then declines as the approaching distance increases. In the proximal zone, the temperature increase is minimal and primarily influenced by in-situ stress. In the intermediate zone, accelerated fracture propagation and continuous gas pressure loading lead to a significant temperature rise. In the distal zone, the temperature increase diminishes due to the predominant effect of gas desorption endothermic behavior.

Considering the heat transfer between the fracturing fluid and the rock, a thermo-hydro-mechanical coupled cohesive phase-field model is proposed to address the dynamic fracture propagation problem in supercritical CO? fracturing. The failure mode of the rock is assumed to be dynamic quasi-brittle fracture, facilitated by the introduction of an inertial term and the traction-separation law. CO? property parameters under varying temperature and pressure conditions are obtained using REFPROP software and integrated with the governing equations for fluid flow and heat transfer through interpolation functions. A two-dimensional numerical model for supercritical CO? fracturing was established to investigate the fracture propagation behavior under different injection rates, viscosities, reservoir horizontal stress differentials, and initial temperatures. The results indicate that: (1) There exists an optimal displacement for supercritical CO? fracturing, maximizing both fracture length and stimulated area; (2) Provided that the wellhead pressure does not exceed the limit, increasing the viscosity of the supercritical CO? fracturing fluid can reduce leak-off and enhance the stimulated reservoir area; (3) The magnitude of the horizontal stress differential has minimal impact on temperature and pressure changes within the formation, yet it significantly influences fracture morphology; (4) The initial temperature of the reservoir primarily affects rock fracture through thermal stress; higher initial temperatures lead to more complex fracture morphologies. By examining the branching patterns of fractures and their fluid pressure responses, the fracturing effects of supercritical CO? under various engineering and geological parameters are elucidated, providing valuable guidance for the design of field supercritical CO? fracturing operations.

Addressing the frequent occurrences of surrounding rock instability and the unclear mechanisms of bolt anchorage failure in complex stress environments necessitates urgent laboratory research on bolt loading response testing under simulated real working conditions. We have independently developed a full-scale tension-shear-torsion test system for bolt-anchored fractured rock masses, based on the principle of active loading tests where rock masses exert forces on bolts. This system effectively simulates and reproduces the complex stress states experienced by bolts under multiple loading paths in underground tunnels. A key innovation of this system is its design, which allows rock masses to crack, shear-slip, and rotate along their fracture planes, thus replicating the mechanical behavior of bolts in situ. Utilizing a servo-controlled, multi-degree-of-freedom coupled loading system, it applies composite loads of tension (±1 000 kN), shear (±2 000 kN), and torsion (±2 000 N•m) simultaneously. This approach effectively addresses the challenge of simulating multi-dimensional stress coupling on bolts at rock mass fracture planes in actual mining engineering. Functional tests and result analyses indicate that the system can stably achieve comprehensive loading and mechanical response acquisition of bolts under combined tension-shear-torsion loads. This capability accurately reflects the bearing characteristics and failure mechanisms of bolts at rock mass fracture planes. It was observed that the rock bolt experiences combined tension-shear loading when the rock mass shears along its fracture surface, with the anchoring shear stiffness of the test rock bolt being less than its tensile stiffness. Additionally, the conditions of the rock mass significantly influence the deformation capacity of the bolts. We propose an evaluation framework centered on total tension-shear resistance and total displacement for assessing bolt anchorage performance. Furthermore, we discuss application strategies for bolt selection based on test results and performance evaluations. The successful development of this test system provides an innovative experimental platform for uncovering the composite loading failure mechanisms of bolts in deep rock masses and establishing evaluation standards for bolt anchorage performance.

Based on a detailed analysis of boundary conditions, twelve types of boundary conditions for slope stability analysis were proposed. A novel method for evaluating slope stability, utilizing the fundamental theories of small deformation, was introduced. The theoretical solutions for stress distributions within the sliding body can be derived when the stress balance differential equation, coordination equation, macroscopic force balance, torque balance, and force boundary conditions for each element are satisfied. Additionally, the stress continuity conditions between elements can also be met. To facilitate comparisons between the proposed theoretical results and those obtained using the finite element method (FEM), the linear elastic Hooke’s law was employed to derive the strain solutions for the sliding body. A point failure criterion was suggested to calibrate the stress state of the sliding body at each point. Using the Daochi Township landslide in Enshizhou, Hubei Province, as a case study, comparisons were made between the results obtained through FEM and those derived from the proposed theoretical method, demonstrating the feasibility of the latter. This proposed method establishes a comparative foundation for current computational results. The theoretical solutions obtained can be presented in Excel format, ensuring clarity and traceability. 

Roadside backfill mining applies gangue cemented rockfill to replace coal pillars, which significantly reduces the loss of coal resources and promotes the source reduction of mine waste. Cemented rockfill is easy to be damaged and deteriorated under the action of overlying strata. It is still a serious challenge to strengthen the performance of cemented rockfill with low cost and high solid waste ratio. In order to solve the fundamental contradiction between the reinforcement of mechanical performances and the cost reduction of filling materials, this paper uses cellulose nanofibers to reinforce cemented rockfill. The macro-meso-micro mechanisms of nano-modified gangue cemented rockfill were investigated. The mechanical compression, acoustic emission monitoring, X-ray diffraction, thermogravimetric, mercury intrusion and microscopic scanning tests were carried out on cemented rockfill materials with different cellulose nanofiber dosages to investigate their strength, deformation and acoustic emission characteristics, hydration products, pore structure and microstructure characteristics. A two media-three interfaces discrete element model was established considering gangue shape parameters, particle size distribution, and interface transition zone between gangue and cemented matrix. The meso-structural characteristics such as force chain network, crack distribution and particle damage after bearing deformation and failure of cemented rockfill materials were discussed. The results show that the compressive strength of cemented rockfill materials increases first and then decreases with the dosage of cellulose nanofibers. The maximum strength enhancement effect induced by the optimal dosage of 0.10% can reach more than 40%, which is attributed to its promotion effect on hydration reaction and the densification of pore structure and interface transition zone. However, it was also found that excessive cellulose nanofibers inhibited the hydration reaction and formed more nanopores. By adjusting the strength ratio of the cementing matrix and the interfacial transition zone to the gangue, the cellulose nanofibers induce the transformation of the main control fracture of the interfacial transition zone to the synergistic control fracture of the gangue bulking-interfacial transition zone. The crack development of the former starts from the interface transition zone, the cemented matrix is seriously damaged, and the gangue damage is small. The crack development of the latter starts from the expansion of the gangue, accompanied by the cracking of the interface transition zone, and the local gangue failure is more serious than cemented matrix.

The weakly cemented overburden in thick coal seams exhibits significant mining disturbance effects, characterized by low rock strength, water-induced disintegration, and highly developed fractures. These factors lead to intense overburden movement and frequent occurrences of high ground pressure. To mitigate hazards at the source of weakly cemented overburden in thick coal seams, this paper uses the 1101 working face of Zhundong No.2 Mine as the engineering background. It employs theoretical analysis, laboratory testing, numerical simulation, and on-site measurements to investigate the macro and microstructural characteristics and strength properties of weakly cemented overburden, the mechanisms generating strong ore pressure in large-height complex faces, and methods for source control. The findings reveal that the weakly cemented overburden rock is primarily soft rock with a strength of less than 30 MPa, severely affected by calcareous erosion fissures. Pores are developed between fine grains, and there is an absence of cemented filler. Following coal seam mining, the intense movement driven by fractures in the weakly cemented overburden results in excessive mining pressure at the working face, frequent hydraulic support failures, and coal wall sloughing. A fracture-containing basic roof structure model was established, elucidating the instability mechanism of the basic roof structure under the combined effects of calcareous erosion fractures, weakly cemented rocks, and water infiltration through fractures. Utilizing field-measured data on calcareous erosion fractures and Discrete Fracture Network (DFN) meshing technology, the ultra-thick sandy mudstone layer overlying the basic roof was reconstructed. Through key block theory, the distribution and proportion of movable blocks within the massive rock stratum were identified. High-proportion, large-sized vertically subsiding blocks rapidly load destabilized basic roof fracture blocks, leading to high-static and dynamic loading phenomena at the working face. An arch-shaped layout for longwall top coal caving mining was proposed, accompanied by a design for the processes of arch face formation. An arch-shaped roof structure model was developed, revealing the stress rotation trajectory induced by mining activities. The arch-shaped working face modifies the structural morphology of the weakly cemented roof and the load transfer path of the overlying strata, enhancing the bearing capacity of the weakly cemented roof, eliminating the influence of calcareous erosion fractures, and reducing the degree of advanced mining-induced stress concentration. The arch-shaped face control technology for high ground pressure has successfully achieved source governance of roof disasters in the working faces of ultra-thick coal seams with weakly cemented overburden. Both support resistance and microseismic energy release from the overlying strata exhibited a decreasing trend, significantly enhancing the effectiveness of surrounding rock control. The research findings can serve as a reference for managing high ground pressure during the safe mining of thick coal seams under similar conditions.

Hydraulic fracturing is predominantly utilized in coal and rock fracturing; however, it increasingly reveals various drawbacks. In recent years, scholarly attention has shifted towards incorporating additives into water or directly employing water-free fracturing fluids. Consequently, the comparative evaluation of fracturing effectiveness using different fluid media has become particularly crucial. This study utilized a self-developed true triaxial fracturing experimental apparatus for multi-field and multi-phase fluids to conduct fracturing experiments on specimens using guar gum aqueous solution, pure water, nitrogen (N?), and supercritical carbon dioxide (CO?) under varying confining pressures. The fracture propagation patterns in specimens fractured with different fluids were analyzed based on multi-dimensional acoustic emission (AE) data, including temporal, frequency, and spatial localization characteristics. The results indicate that the AE ring count in specimens fractured with liquid media exhibited cyclic fluctuations, whereas specimens fractured with non-liquid media showed a sharp surge in ring-down count near the fracture initiation pressure. The proportion of shear fractures in specimens fractured with guar gum solution consistently ranged from 50% to 60%, demonstrating little difference from tensile fractures. Conversely, specimens fractured with other fluids were predominantly characterized by tensile fractures, often coexisting with mixed tensile-shear fractures. The -value for specimens fractured with liquid media hovered around 1.0, even exhibiting a decreasing trend. In contrast, specimens fractured with nitrogen and CO? displayed an inflection point in the -value, rising sharply near the initiation pressure, indicating significant energy dissipation and specimen instability in the later experimental stages, which led to the formation of complex fracture networks. The dominant AE frequencies in the time domain for specimens fractured with guar gum solution and pure water were primarily distributed within the ranges of 200–400 kHz and ≥800 kHz. In contrast, the dominant frequencies for specimens fractured with nitrogen and CO? were concentrated entirely below 400 kHz, exhibiting a “multi-spectral-band” signal pattern. Following the achievement of initiation pressure, the occurrence of AE-located events began to increase rapidly, coinciding with the emergence of high-energy events for all fluids. The propagation and coalescence of micro-fractures ultimately resulted in the formation of macroscopic fractures. The fracturing process with liquid media was relatively moderate, whereas specimens fractured with nitrogen and CO? developed complex fracture networks. This phenomenon is attributed to the instantaneous fracturing resulting from the high-pressure accumulation of non-liquid media within the specimens.

The Cone Penetration Test (CPT) is one of the most important in-situ methods for geotechnical site investigation. To address issues encountered in engineering practice, such as localized data gaps and significant variations in CPT measurements due to probe malfunctions and abrupt changes in soil properties at stratigraphic interfaces, this paper proposes a physics-Bounded Multi-Task Bayesian Compressive Sensing (B-MTBCS) methodology. This approach establishes an information fusion mechanism for multiple CPT soundings within a Bayesian framework. By incorporating physical constraints—specifically, the non-negativity of CPT responses—as boundary conditions for predicted data, it allows for accurate estimation of missing CPT measurements. Validation through numerical simulations and real-world engineering case studies demonstrates the method’s effectiveness in recovering missing CPT data under complex geological conditions. Compared to the conventional Multi-Task Bayesian Compressive Sensing (MTBCS) approach, the proposed method reduces prediction errors by over 34%. These research findings provide substantial support for critical geotechnical applications, including parameter inversion, foundation bearing capacity evaluation, and liquefaction potential assessment in sandy soils.

The shear strength of rock joints is a critical parameter for the stability analysis of rock engineering, and the use of experimental instruments is essential for accurately determining this parameter. Due to variations in loading methods, constraints, and control strategies among current shear strength testers, discrepancies can arise in test results obtained from the same rock sample. Based on the testing principles outlined in shear strength testing standards, technologies such as normal load-following loading, tangential centering loading, and multimodal force-position conversion control have been proposed, leading to the development of an innovative rock joint shear strength tester. To validate its accuracy and reliability, an experimental study was conducted on the shear strength of a series of flat and undulating structural samples. The results indicated that the tester effectively achieves real-time centering loading under conditions of significant freedom during the slip process of the sample, while exhibiting minimal frictional resistance in the loading mechanisms. Compared to the conventional “upper shear” test method, the shear load curve obtained is smoother and yields more reasonable results. In comparison to the traditional “lower shear” test method, the shear load curve more accurately reflects the influence of structural surface morphology. The multimodal force-position conversion control method simulates the entire process of stress accumulation, release, and maintenance during direct shear tests at various stages, allowing for the precise acquisition of peak shear strength and stable test data, in contrast to the tangential constant rate displacement control method. The findings of this research provide theoretical and technical support for experimental investigations and the development of related testing instruments.

Inspired by the fundamental principle that inherent defects, such as primary pores and fractures in natural rocks, significantly influence their mechanical properties, this paper introduces the Mechanic-Controlled Sampling via Structure-Driven Design (MSSD) method for fabricating transparent rock-like samples with controllable mechanical characteristics. Utilizing this method, we fabricated intact and pre-fissured transparent rock-like samples with varying porosities (0%, 3%, 6%, 10%, 20%, and 30%). We established quantitative relationships between porosity and key mechanical parameters, including uniaxial compressive strength, elastic modulus, tensile strength, brittleness, cohesion, and internal friction angle. The influence mechanism of pores on crack propagation was elucidated through Franc3D simulations. Experimental results from intact samples demonstrate that the failure modes are highly consistent with those observed in natural rocks, allowing for clear observation and capture of the crack propagation process. As porosity increases, the various mechanical parameters of intact samples exhibit a linear decreasing trend, transitioning from hard rock-like material to soft rock-like material. Experimental results from pre-fissured samples indicate that increasing porosity leads to heightened surface roughness of wing cracks in single-fissure samples. In double-fissure samples, the rock bridge coalescence mode shifts from global to local coalescence, with the stress intensity factor KIII at the crack tip gradually decreasing, while the fracture mode in the rock bridge area transitions from tensile to shear type. Moreover, an increase in porosity correlates with a reduction in the number of petal-shaped cracks observed in all pre-fissured samples.

To elucidate the influence of the coupling effect of a water-rich temperature environment and time on the mechanical properties of the grouted rock mass, we enhanced a self-constructed three-dimensional rock mass grouting simulation test device. Employing a uniform design method, five working conditions were established for conducting high-pressure grouting tests. A total of 270 standard cylindrical samples were prepared, with multi-gradient curing times ranging from 7 to 90 days in water-rich temperature environments, enabling a systematic investigation of the time-dependent evolution of compressive strength and permeability characteristics of the grouted rock mass. The findings demonstrate that the improved three-dimensional test system can accurately simulate the grouting environment of 20 ℃ to 60 ℃surrounding rock. Furthermore, a multi-field coupling physical test platform was successfully constructed, integrating temperature, seepage, and stress fields. The mechanical properties of the grouted rock mass exhibited a pronounced temperature threshold response. Specifically, the compressive strength in a 50 ℃ water-rich environment displayed a three-stage evolution characterized by peak-attenuation-stability over curing time, aligning with the Gaussian distribution law. After 28 days of curing, the influence of the water temperature factor surpassed that of the water-cement ratio and grouting pressure. Additionally, a double-critical transition phenomenon was noted in the permeability characteristics. The permeability coefficient in the 50 ℃ water-rich environment adhered to a cubic polynomial trend, while in the 60 ℃ water-rich environment, the 28-day curing time served as a turning point, resulting in abnormal permeability changes in the later stages, transitioning from continuous reduction to linear growth. Under the 60 ℃ water-rich conditions, a significant negative correlation was observed between compressive strength and permeability, with the two trend curves exhibiting two cross-turns. This observation revealed the deterioration mechanism linked to thermal damage accumulation and micro-fracture network evolution in the grouted rock mass due to the synergistic effects of water-rich temperature, seepage, and time. By transcending the traditional limitation of a 28-day curing period, the derived evolution law regarding the performance of the grouted rock mass, based on the dynamic water-rich temperature-time threshold, provides a theoretical foundation for optimizing grouting times in deep-buried tunnels and ensuring the long-term stability of surrounding rock.

To investigate the characteristics of rockburst failure and the intensity levels of rockbursts in deep-buried roadways with straight walls and structural planes, a true-triaxial test was conducted on cubic limestone samples extracted from indoor arched roadways featuring straight walls and structural planes. The failure of the rock mass was monitored in real-time using a miniature camera and acoustic emission techniques. The stress coefficient for the rock mass was defined using Fish language and subsequently imported into FLAC3D numerical simulation software to predict the rockburst intensity levels of roadways with concealed straight structural planes. The research findings indicate that the straight-wall arch-shaped specimen without exposed straight structural planes experiences rockburst failure, whereas the specimen with exposed straight structural planes undergoes rock slab cracking failure. Furthermore, the extent of damage and the depth of the “V” groove failure zone in the former specimen is greater than that in the latter. The cumulative ringing count and absolute energy of acoustic emissions from the unexposed flat structure test sample exceed those of the exposed straight structure test sample. Notably, the proportion of shear failure in the unexposed straight structure test sample during the entire loading process is significantly lower compared to that of the exposed straight structure sample. The presence of undisclosed straight structural planes leads to variations in the rockburst intensity levels across different sections of the roadway. Specifically, the rockburst intensity level of the side walls containing undisclosed straight structural planes has increased from level 1 to level 2. These research results offer new insights for the prevention and control of rockbursts in underground caverns subjected to high stress and dynamic disturbances.

To investigate the thermal evolution of surrounding rock in high-water-temperature tunnels where both heat conduction and convective heat transfer between geothermal water and rock coexist, this study derives a finite difference solution for the coupled thermal-hydraulic temperature field and develops a three-dimensional numerical model that integrates the high-temperature water-saturated rock mass with tunnel air. A parametric study is conducted to analyze the effectiveness of grouting for water sealing and thermal insulation. Additionally, a physical model test system for high-temperature tunnels is developed to validate the finite difference solution and numerical results. The findings indicate that: (1) when rock permeability exceeds 5×10?? m/s, convective heat transfer from thermal water maintains the rock at its initial high temperature; as permeability decreases, rock temperature declines, but further reduction below 1×10?? m/s yields diminishing cooling returns. (2) Grouting significantly reduces convective heat transfer between water and rock, thereby lowering the temperatures of both rock and air. For rock with a permeability of 1×10?? m/s, a grout zone with permeability two orders of magnitude lower and a thickness of 3 m achieves more pronounced cooling. (3) Model tests demonstrate that initial ventilation leads to rapid rock cooling, which stabilizes over time; wall temperature trends align with the finite difference solution, and the relative errors in wall and outlet air temperatures between simulation and experiment are within 15% and 5%, respectively. Combined grouting and enhanced ventilation are recommended for thermal control in high-temperature tunnels. The results provide theoretical guidance for the cooling design of similar geothermal tunnels.

Currently, research on slow-moving slopes is hindered by a lack of integration between the temporal and spatial dimensions in analyzing movement characteristics and their underlying causes. Additionally, the potential instability of the No. 3 slope at Wangjiapo in Bailuyuan, an important residential and tourist area, poses significant risks of casualties and economic losses. To investigate the movement mechanism of this slow slope, we developed a collaborative analytical framework consisting of “temporal InSAR monitoring, multi-scale signal decomposition, three-dimensional numerical simulation, and field verification.” Initially, we extracted the annual average velocity field and time sequence displacement of the landslide mass movement by comprehensively utilizing SBAS-InSAR and StaMPS-InSAR technologies. Furthermore, we decomposed the high-precision InSAR time series displacement into periodic and trend components using wavelet decomposition. The trend component was fitted using a multi-segment nonlinear fitting method to calculate the temporal movement velocity of the landslide. We conducted Pearson correlation analysis between deep displacement measurements, GNSS data, and the wavelet-decomposed trend displacement. The relationship between rainfall and the displacement changes of the periodic components was quantified through linear regression combined with correlation analysis. Subsequently, we employed three MESH finite difference numerical models with varying degrees of freedom to assess the stability of the landslide and simulate its motion characteristics using the intensity reduction method. Finally, we performed field verification based on the velocity field derived from time-series InSAR, the wavelet-decomposed time-series displacement, and the numerical simulation results. The findings indicate that: (1) StaMPS-InSAR demonstrates higher accuracy compared to SBAS-InSAR. The InSAR velocity field reveals that the upper portion of the landslide moves more slowly, while the middle and lower sections exhibit faster velocities. (2) The Pearson correlation coefficients between the wavelet-decomposed InSAR trend displacement and the deep displacement measurements and GNSS data were 0.96 and 0.95, respectively. (3) The calculated time series velocity, following multi-segment nonlinear fitting of the trend displacement, shows spatiotemporal variations in the landslide’s movement, with fewer acceleration events in the upper part compared to the middle and lower sections. This variation in velocity changes correlates with differences in soil strength at different locations. Notably, there is no significant linear correlation between the periodic displacement changes and rainfall. The safety factor obtained from numerical simulations was 0.967, indicating that the lower part of the slope is a stress concentration area, consistent with findings from InSAR and on-site investigations. This analytical framework has achieved a multi-dimensional spatiotemporal analysis of landslide movement through the organic coupling and progressive verification of multiple technical methods, providing a scientific basis for addressing the Wangjiapo No. 3 landslide and offering valuable insights for studying the movement mechanisms of similar slow-moving landslides.

To characterize the influence of micro- and meso-structures on the macroscopic mechanical properties of reef limestone, this study investigates two typical types of reef limestone. Through multi-scale physical and mechanical experiments, the intrinsic correlations among microscopic minerals, mesostructural characteristics, and macroscopic mechanical properties are systematically elucidated. Utilizing digital rock technology and mesoscopic homogenization methods, a multi-scale statistical damage constitutive model for reef limestone is established. The experimental results reveal significant variability in the macroscopic mechanical properties of reef limestone. The two types of reef limestone exhibit distinct differences in pore structure at the mesoscopic scale, while nanoindentation and X-ray diffraction (XRD) analyses indicate that their microscopic mineral compositions and matrix mechanical properties are highly similar. Consequently, the pore structure is identified as the key factor influencing the macroscopic mechanical response of reef limestone. By employing statistical damage theory and digital rock technology, the reef limestone is conceptualized as an assemblage of microelements forming a two-phase pore-matrix system. Assuming that microelement failure adheres to the Drucker-Prager strength criterion, the equivalent strength of the microelements is derived using the established microelement equivalent strength model. The strength probability distribution function of the microelement assemblage is then fitted to construct the multi-scale statistical damage constitutive model for reef limestone. Validation results demonstrate that the proposed model effectively captures the entire deformation process of reef limestone, from linear elasticity to peak strength, enabling cross-scale characterization from microscopic heterogeneity to macroscopic mechanical behavior. Although the model has certain limitations in representing the deformation mechanisms during the compaction stage, it provides a theoretical foundation for understanding the mechanical behavior of porous rocks.

To address the limitations of traditional fracture phase-field models—including their challenges in accurately describing tension-shear mixed-mode cracks in rock, their heavy reliance on meshes within mechanical solution frameworks, and their susceptibility to the “curse of dimensionality” in high-dimensional problems—this paper proposes a novel method that integrates Physics-Informed Neural Networks (PINNs) with a mixed-mode phase-field fracture model to simulate the tension-shear failure behavior of rock under loading. First, based on the modified fracture F-criterion, the method decomposes elastic energy into tensile and shear strain energy components through volumetric-deviatoric splitting. By combining this with the critical energy release rates for tension and shear, it derives the governing equations of the mixed-mode phase-field model for tension-shear cracks. Second, the study employs PINNs to construct an adaptive solution framework and optimizes the energy functional using the Deep Ritz Method (DRM), thereby circumventing the optimization conflicts typically associated with traditional residual-based loss functions. Validation through numerical examples—including the single-edge notched tension test, inclined crack propagation test, and parallel crack staggered propagation test—demonstrates that the calculated results align well with laboratory test data, confirming the correctness and effectiveness of the proposed method. Additionally, by varying the phase-field characteristic length, the regulatory mechanism of crack diffusion effects on fracture behavior is elucidated. This study not only provides a new theoretical framework and numerical implementation pathway for addressing rock mixed-mode failure issues but also establishes a foundation for researching real-time predictions of rock mechanical responses under varying initial conditions, including boundary conditions and material properties, by integrating PINNs with high-fidelity datasets.

The negative coal pillar roadway arrangement has been widely adopted in rockburst-prone mines across the country, demonstrating effectiveness in mitigating the rockburst risk associated with roadways along the gob. However, an unexpected rockburst incident occurred in the negative coal pillar roadway during the mining of the 17-seam coal gob-side fully mechanized caving face at Junde Coal Mine. In response to this incident, we conducted a comprehensive investigation of the rockburst damage characteristics of the negative coal pillar roadway in the context of dynamic load disturbances. Our approach utilized a multifaceted methodology, incorporating field measurements, numerical simulations of particle flow, and theoretical analysis. This enabled us to clarify the underlying mechanisms of rockbursts in negative coal pillar roadways induced by dynamic load disturbances and to propose a methodology for identifying rockburst hazards in such roadways. The findings of this study can be summarized as follows: We analyzed the movement and evolution characteristics of surrounding rock in the negative coal pillar roadway and proposed a risk criterion for rockbursts under dynamic load disturbances based on the ejection velocity of coal and rock bodies and the maximum radial displacement of the roadway. A rockburst damage calculation model for the negative coal pillar roadway under dynamic load disturbances was developed based on energy attenuation theory and rock dynamics. We derived the displacement damage engineering criterion and the dynamic damage criterion for the coal body in the negative coal pillar roadway under dynamic load disturbances, and proposed a method for identifying rockburst risk in this context. From a displacement perspective, the radial displacement of the roadway depends on the stress of the original rock, the angle of internal friction, and cohesion, which reflect the inherent strength of the rock, support resistance, and dynamic load. From an energy perspective, the ejection velocity of the coal body in the roadway is influenced by the energy of the mining shock, the distance from the mining shock to the negative coal pillar roadway, the energy attenuation coefficient, the range of the plastic zone, and the energy-absorbing effect of the roadway support assemblies. We proposed three targeted prevention and control measures, focusing on the generation of the vibration source, the propagation path of the vibration wave, and the mechanism of vibration wave disaster. These measures are designed to reduce the risk of rockbursts in negative coal pillar roadways subjected to dynamic load disturbances. The first measure involves quantifying the critical width of the pressure relief protection zone. The second measure addresses the over-support of the negative coal pillar roadway. The third measure entails pre-cracking and blasting of the roof plate.

Currently, there is a considerable body of research focused on the thermo-mechanical behavior of energy piles under static loading; however, studies investigating the dynamic response of energy piles under cyclic loading remain limited. This study examines the dynamic response of single energy piles in sandy soil through model testing, analyzing changes in pile top displacement and cyclic stiffness characteristics. Additionally, a three-dimensional finite element model that accounts for the relevant behavior of the pile-soil contact surface under cyclic shear was developed and validated against the test results. Using this numerical model, the effects of the number of cycles, cyclic loading frequency, and temperature gradient on the vertical dynamic response of single energy piles were assessed. The findings indicate that as the number of cycles increases, the pile tip resistance appears to strengthen while the pile side resistance weakens, with the variation trend under heating conditions being more pronounced than that under cooling conditions. Furthermore, an increase in cyclic loading frequency leads to greater cumulative settlement of the energy pile, with the pile tip resistance showing an upward trend and the average pile side resistance a downward trend. Notably, the strengthening of pile tip resistance and the weakening of pile side resistance become less pronounced once a certain threshold is reached.

To investigate the moisture migration and immersion-induced deformation characteristics of loess high-fill foundations, a field immersion test was conducted on a layered compacted high-fill project in Lanzhou. Over a period of 364 d, the volumetric water content, surface settlement deformation, and crack development around the test pit were monitored. This study analyzed the moisture migration patterns, variations in vertical and horizontal permeability coefficients, and the wetting deformation characteristics of the fill foundation. The results indicate the following: (1) During immersion, daily water consumption exhibited distinct stage-wise characteristics. The upper section of the fill foundation can enhance anti-seepage effectiveness by employing loess mixed fill and increasing compaction. (2) Evaporation must be accounted for in field immersion tests. Daily infiltration stabilized at 2.5–3.0 m3/d, with evaporation constituting approximately 2/3 to 1/2 of the daily water consumption. (3) Full development of wetting deformation requires multiple cycles. As fill depth increases, the onset of deep soil wetting is delayed, resulting in long-term post-construction settlement. (4) The vertical permeability rate decreases with increasing fill depth, while the horizontal permeability rate remains consistent, leading to an oblate elliptical moisture diffusion pattern. (5) The “boundary effect” of immersion deformation in fill foundations differs significantly from that in undisturbed loess foundations. For fill foundations located near water sources, a minimum safety distance of 10 m should be maintained between structures and the water source. (6) During the immersion process of the fill foundation, the surface settlement rate in the last month should not exceed 2.5 mm/d as the criterion for ceasing immersion. These findings provide a scientific basis for the treatment of fill foundations and the mitigation of wetting deformation issues in loess regions.