Restricted by the significant impact of the “three-high” geological environment (high ground stress, high osmotic pressure, and high ground temperature) on the quality of surrounding rock and the construction process, current research faces three core challenges: (1) Traditional surrounding rock classification methods are often ill-suited for high-energy geological environments; (2) The mechanism of the “three-high” coupling effect on the mechanical properties of surrounding rock is complex and difficult to elucidate; (3) The accuracy of obtaining in-situ surrounding rock parameters in deep regions is inadequate. This paper summarizes the current research status and developmental trends of surrounding rock classification and parameter acquisition methods under the “three-high” geological environment, and proposes directions for future research and development. Overall, research on surrounding rock classification and parameter acquisition methods is progressively moving towards integration, dynamization, and intellectualization, with key developmental trends manifested in the following aspects: (1) There is an urgent need to establish a multi-field coupling surrounding rock classification system that is adapted to the “three-high” geological environment. The coupling effect of the “three-high” geological environment significantly influences the mechanical properties, deformation characteristics, and permeability of surrounding rock, directly affecting the establishment of classification systems tailored for various construction methods. It is essential to integrate geomechanical parameters with real-time monitoring data and employ machine learning technologies to accurately assess the stability, excavability, and construction risks of surrounding rock, thereby meeting the construction requirements of both the drilling and blasting method and the TBM method. (2) There is an urgent need to develop dynamic methods for determining surrounding rock mechanical parameters. Methods for determining these parameters in high-energy geological environments must fully consider the dynamic feedback mechanisms and actively incorporate advanced technologies such as deep learning and intelligent identification to update classification results and construction strategies in real-time. (3) There is an urgent need to construct a multi-field coupling parameter acquisition model. Efforts should focus on transitioning from empirical judgment to theoretical models, establishing constitutive relationships for obtaining surrounding rock parameters based on high-energy geological environments, and further developing cross-scale and multi-source information fusion methods to improve parameter accuracy. In the future, it is crucial to address the bottlenecks in in-situ testing and dynamic feedback technologies for deep surrounding rock, and to establish classification standards for surrounding rock that are suited to high-energy geological environments, thereby providing theoretical support and technical guarantees for the safe construction of deep engineering projects.

Under periodic mining disturbances, the surrounding rock masses of roadways in deep engineering are subjected to cyclic loading conditions, which promote cumulative fatigue damage and plastic deformation, ultimately compromising the long-term stability of the roadway. To elucidate the mechanical behavior and multi-source responses of muddy sandstone from the roadway roof under cyclic loading, a series of laboratory uniaxial cyclic loading-unloading tests were conducted to systematically investigate the effects of the upper stress level (15.6–39.0 MPa), lower stress level (0–23.4 MPa), and cycle number (10–40 cycles) on fatigue degradation. Synchronous monitoring of “mechanical parameters–acoustic emission (AE)-P-wave velocity-apparent resistivity-displacement field” was implemented to establish a mutually corroborative, multi-source characterization framework. The results indicate that increases in the upper stress level, lower stress level, and cycle number lead to changes in post-cycling peak stress ( ) of －22.82%, +36.55%, and －36.01%, respectively, and changes in peak strain ( ) of +14.94%, +21.35%, and －17.77%, respectively. Multi-source observations further reveal that the elastic strain ratio per cycle (Re) exhibits an exponential relationship with the cycle number (R2 = 0.805 to 0.962), reflecting an evolution from initial plastic damage to internal structural reorganization and an increased elastic proportion. AE activity demonstrates a staged pattern of “high activity in early cycles-quiescence in mid-to-late cycles-reactivation after cycling-sharp surge near pre-peak yielding.” The P-wave velocity evolves through a sequence of “initial compaction-induced increase-progressive damage-induced decrease-abrupt drop near pre-peak yielding.” Apparent resistivity exhibits a pattern of “localized increase during cyclic loading-partial decrease during unloading-pronounced high-resistivity band near the peak-gradual homogenization after multiple cycles.” Displacement-field measurements reveal marked gradient contrasts during cyclic loading that diminish after unloading, with final failure characterized by tensile splitting exhibiting mixed tensile-shear characteristics. Collectively, the concurrent occurrence of “pre-peak velocity drop + synchronous AE surge + intensified high-resistivity zone + pronounced displacement-gradient contrast” indicates that the rock has entered a critical transition from stable damage accumulation to unstable failure, providing a practical reference criterion for early warning of instability in surrounding rock subjected to cyclic disturbances in situ.

To systematically investigate the evolution patterns and intrinsic relationships between the micro-mesoscopic structure and macroscopic mechanical properties of open-pit mine mudstone under varying temperatures, samples from an open-pit mine in Inner Mongolia were subjected to heat treatment at temperatures ranging from 25 °C to 600 °C. Multiscale characterization techniques, including X-ray diffraction (XRD), computed tomography (CT) scanning, and scanning electron microscopy (SEM), were employed to analyze the evolution of mineral composition, crystallinity, pore structure, and microcrack distribution. In conjunction with uniaxial and triaxial compression tests, the influence of temperature on mechanical parameters such as strength, elastic modulus, and Poisson’s ratio was examined, elucidating the restraining effect of confining pressure on high-temperature damage. The main findings are as follows: (1) While mineral composition remained relatively stable, crystallinity demonstrated a nonlinear trend, initially increasing and then decreasing, with a peak of 77.9% observed at 400 °C. (2) Porosity increased from 6.8% to 17.9% with rising temperature, as the pore structure evolved from smaller to larger pore sizes; additionally, microcracks initiated, propagated, and interconnected, ultimately forming a highly connected three-dimensional fracture network at 600 °C. (3) High-temperature exposure shifted the failure mode from ductile to brittle. The peak strength under uniaxial compression and crystallinity exhibited a consistent trend with temperature variation, indicating that microstructural evolution is a key factor governing macroscopic mechanical behavior. (4) Under varying confining pressures, both peak strength and elastic modulus of mudstone increased with rising temperature, while the effectiveness of confining pressure in mitigating thermal damage improved with increased confining pressure. This finding suggests that confining pressure can inhibit the propagation of microcracks and partially offset the thermal damage caused by high temperatures. The research results provide a crucial basis for developing slope control measures in open-pit coal mines under conditions of spontaneous coal combustion.

Deep underground roadways with surrounding rock that has developed embedded fractures are susceptible to deterioration under frequent and intense dynamic loads, resulting in engineering disasters. To investigate the influence of randomly distributed embedded open fractures on the mechanical characteristics and damage evolution of rock masses under dynamic loading, this study prepared rock-like specimens with varying numbers of randomly distributed embedded open fractures using sand-powder 3D printing technology. Cyclic disturbance loading tests were conducted, and acoustic emission (AE) monitoring technology was employed to analyze the differences in strength and deformation characteristics, AE behaviors, failure behaviors, and energy characteristics of the rock-like specimens with varying numbers of fractures. The damage evolution characteristics of the different specimens were analyzed based on dissipative energy density and AE energy. The results indicate that: (1) The strength of specimens with varying numbers of embedded fractures varies significantly and shows a negative correlation with the number of fractures. The deterioration performance of different specimens during the dynamic load segment differs, with both the average decline rate of the loading modulus and the average growth rate of cumulative irreversible deformation negatively correlated with the number of fractures. (2) Specimens consistently exhibit intensified damage characteristics during the dynamic load segment. The presence of embedded fractures increases the proportion of shear-type cracks, and a higher number of fractures leads to an earlier onset of macroscopic failure. Moreover, the failure paths and final failure modes are significantly influenced by the number of fractures. (3) The average energy dissipation rate during the dynamic load segment exhibits a positive correlation with the number of fractures. Both the proportion of damage during the dynamic load segment and the average damage rate show positive correlations with the number of fractures. The research findings can provide a reference for conducting mechanical tests on fractured rock masses using sand-powder 3D printing technology and serve as a foundational basis for the stability control of engineering fractured rock masses.

To investigate the effects of freeze-thaw (F-T) cycles on the mechanical response and permeability evolution of highly weathered granite under seepage-stress coupling conditions, nuclear magnetic resonance (NMR) tests and seepage-stress coupled triaxial compression tests were conducted on highly weathered granite specimens subjected to varying numbers of freeze-thaw cycles. The influences of freeze-thaw cycles and seepage pressure on the pore structure, strength, deformation, failure modes, and permeability characteristics of the rock were examined. The results indicate that freeze-thaw cycles lead to a slight decrease in rock mass, a significant increase in saturated water content (with a 46% increase after 50 cycles compared to the initial state), and a considerable rise in porosity, resulting in a more complex pore structure. As the number of freeze-thaw cycles and seepage pressure increases, the peak stress, elastic modulus, and axial strain stiffness decrease markedly, while the dilatancy stress and the corresponding volumetric strain also exhibit a declining trend. The number of freeze-thaw cycles has a pronounced effect on the failure mode: unfrozen specimens demonstrate clear brittle tensile cracking on the surface, whereas freeze-thaw-treated specimens primarily exhibit shear failure. Furthermore, permeability increases significantly with the number of freeze-thaw cycles and pore pressure. Freeze-thaw damage amplifies the sensitivity of initial permeability to seepage effects, while the maximum permeability is predominantly influenced by the rock failure mode. These findings provide experimental evidence and parameter references for the durability assessment and seepage stability analysis of dam-foundation rock masses in pumped-storage hydropower reservoirs located in cold regions.

Rock size is a critical factor influencing the macro-meso mechanical properties of soil-rock mixtures (SRM). To investigate the effect of particle size, this study conducted large-scale triaxial shear tests on SRM using the GDS apparatus, maintaining a constant relative density. Tests were performed on SRM with six different rock sizes under three confining pressures. The results indicate that the stress-strain behavior of SRM primarily exhibits strain softening. The variation of shear strength and initial elastic modulus with respect to Rd (the ratio of specimen diameter to the maximum particle size) can all be described by a power function. Rd = 10 is identified as the critical ratio; when Rd≥ 10.0, both the shear strength and strength parameters stabilize. The failure mode of SRM is jointly influenced by Rd and confining pressure. At a confining pressure of 100 kPa, the failure mode is characterized by shear failure. At 200 kPa, the failure mode transitions from shear failure to bulging failure as Rd increases. At a confining pressure of 400 kPa, the failure mode manifests as bulging failure. Under the same confining pressure, the particle breakage ratio follows a decreasing power-law trend with increasing Rd. These research findings are significant for accurately predicting the macroscopic mechanical properties of SRM.

Deep underground engineering often occurs in environments with high ground temperatures, which significantly impact the diffusion characteristics of grout in fractured rock masses. Understanding the diffusion mechanism of grout in these conditions is crucial for the design of grouting reinforcement parameters in deep high-temperature engineering projects. This study utilizes the Low-Field Nuclear Magnetic Resonance (LF-NMR) monitoring system to conduct Bingham-type grout diffusion experiments under varying temperature conditions. Real-time observations and analyses were performed on the diffusion characteristics—including grouting volume evolution, permeability, and relative filling degree—of grout in porous fractured sandstone at different temperatures.
Furthermore, using the Finite-Discrete Element Method (FDEM) numerical simulation, this research investigated the effects of temperature on the diffusion characteristics (diffusion distance, grouting pressure evolution, and flow rate) of Bingham-type grout in porous fractured sandstone, as well as the evolution of fracture aperture. By integrating experimental and numerical simulation results, we elucidated the diffusion mechanism of grout in fractured rock masses at elevated ground temperatures. The findings reveal that high temperatures accelerate the hydration process of grout, resulting in increased plastic viscosity and significantly inhibiting the diffusion performance of the grout. At 80 °C, the grouting volume decreased by 40.5% compared to 20 °C, while the permeability of both the effective grouting section and the overall section declined by 33.7% and 31.1%, respectively. Higher temperatures correspond to shorter grout diffusion distances over the same time period and greater grouting pressure requirements for achieving similar diffusion distances (the grouting pressure in the hole increased by 70% at 80 °C compared to 20 °C). This phenomenon is attributed to the increased hydrolysis reaction rate of ethylene glycol diacetate at elevated temperatures, which generates more acetic acid, accelerates its polymerization reaction with sodium silicate, and produces additional gel. This process raises the plastic viscosity of the grout and enhances the shear interaction between the grout and the fracture walls, significantly increasing the grouting pressure necessary to advance the grout front. Under these conditions, the effective diffusion distance of grout in fractures is markedly shorter than at room temperature. With the continuous application of grouting pressure, the grout at the expansion front tends to penetrate into the surrounding rock matrix, continuously consuming the grout’s kinetic energy and thereby diminishing its capacity to further promote front expansion. The results of this research provide a theoretical foundation and technical reference for grouting engineering in high-temperature strata.

The wave velocity characteristics of rock materials comprehensively reflect their physical properties, macro- and microstructural features, and stress states, thereby providing a fundamental basis for identifying rock mass structures and mechanical behavior using acoustic techniques. To elucidate and quantitatively characterize the relationship between stress and wave velocity in rocks, this study integrates dynamic analysis of forced particle vibration with acoustoelastic theory to clarify the evolution mechanism of particle vibration characteristics under static loading. A functional model correlating rock stress and wave velocity is established by introducing the tangential elastic modulus and the dynamic elastic modulus as intermediate variables. This approach reveals the mechanism through which stress influences wave velocity and enables a quantitative description of wave velocity responses to changes in stress state. Theoretical derivations demonstrate that static force alters only the equilibrium position of particle oscillations within a material, without affecting the intrinsic vibration characteristics induced by acoustic excitation. Stress does not explicitly appear in the incremental vibration equation or the wave velocity formulation; rather, it indirectly influences wave velocity by modifying the material stiffness parameters. This characteristic constitutes a fundamental distinction between acoustic wave propagation in rocks and that in ideal isotropic materials. Experimental results indicate that, with increasing stress, rock wave velocity evolves through four distinct stages: an initial gradual increase, a rapid increase at intermediate stress levels, minor fluctuations prior to peak stress, and a rapid decrease post-failure. Within the linear elastic regime of the stress-strain response, the tangential elastic modulus varies synchronously with wave velocity and exhibits a strong linear correlation with the dynamic elastic modulus governing wave propagation. Based on the experimental results and critical stress points, a mathematical model describing the stress-wave velocity relationship is derived, clarifying the physical significance of the model parameters. In conjunction with existing studies, the applicability and limitations of the proposed model are further discussed. This study aims to reveal the microscopic driving mechanisms by which stress alters wave velocity, providing a quantifiable stress-wave velocity relationship for rock mechanics research and offering a scientific basis for stress state identification in rock engineering applications based on acoustic monitoring techniques.

Slope deformation stability is a key concern in slope engineering. Unlike the safety control standards for anti-sliding stability, the complex and variable nature of slope evolution has so far prevented the establishment of a unified criterion to quantitatively determine whether deformation has converged to a stable state. To address this, a database of 650 slope cases is compiled and classified into three categories: deformation-convergent (145 cases), potentially unstable (340 cases), and failure (165 cases). Monthly (?Sm) and annual (?Sa) displacement increments during both stable and secondary deformation stages are statistically analyzed and compared. Risk characteristics of ?Sm and ?Sa are further examined using risk measurement indicators. The results indicate that: (1) For deformation-convergent cases, the mean ?Sm and ?Sa are approximately 0.04 ± 0.03 mm/month and 0.79 ± 1.29 mm/year, respectively. The vast majority of cases exhibit deformation rates below 0.1 mm/month (≈ 91.9%) and 1 mm/year (≈ 78.3%). In contrast, potentially unstable and failure cases display significantly larger ?Sm and ?Sa values, with data points dispersed across multiple orders of magnitude, showing clear differentiation from the deformation-convergent cases. (2) Based on statistical and risk analyses, thresholds of ?Sm = 0.1 mm/month and ?Sa = 1 mm/year can serve as unified convergence stability criterion. When ?Sm≤0.1 mm/month and ?Sa≤1 mm/year, slopes can be considered stable. (3) The reliability of this criterion is further validated through three representative cases of the toppling deforestation slope of Longtan Hydropower Station (stable), the left-bank slope of Jinping-I Hydropower Station (partially stable), and the Xintan landslide (unstable). The results provide a practical and scientific basis for slope stability assessment, risk early warning, and engineering acceptance.

Accurately and efficiently characterizing the mechanical response of surrounding rock is fundamental to ensuring safe tunnel excavation. To address the challenges associated with conventional analytical and numerical methods—such as complex model construction, limited generalizability, and difficulties in balancing computational efficiency with accuracy—this paper proposes a dual data- and physics-driven approach for analyzing the mechanical response of tunnel surrounding rock. First, based on the secondary stress superposition principle and the Verruijt-Booker governing equations, we construct a neural network approximation that outputs displacement and stress fields. This network architecture inherently embeds the governing equations, boundary conditions, and initial conditions, resulting in a mesh-free Physics-Informed Neural Network (PINN). During training, we optimize millions of network parameters to identify a high-dimensional function that simultaneously satisfies physical laws and data constraints, ultimately forming an intelligent surrogate solver for the system of partial differential equations. Second, to tackle issues such as highly variable geological conditions, diverse design scenarios, and sparse monitoring data, we introduce a Transfer Learning-Driven Physics-Informed Neural Network (TL-PINNs). By employing a pre-training and parameter transfer strategy, the model achieves rapid convergence within a high-dimensional parameter space involving multiple variables, including tunnel depth, diameter, and rock mass properties. Results demonstrate that the PINN model closely reproduces outcomes from classical Finite Element Method (FEM) simulations. Under constraints from 45 monitoring points and 6,000 Latin Hypercube Sampling (LHS) collocation points, the maximum absolute errors in horizontal and vertical displacements are 0.31 mm and 0.23 mm, respectively. Compared to the baseline PINN without transfer learning, the TL-PINNs model reaches the same loss level in only 5 000 iterations—a fraction of the 50 000 required by the conventional approach—while reducing training time from 3 664 seconds to 420 seconds, achieving up to a 66.1% reduction in absolute displacement error. As a mesh-free intelligent algorithm, the TL-PINNs framework offers high computational speed, strong interactivity, and excellent generalization capability. The proposed dual data- and physics-driven framework meets practical engineering requirements in terms of accuracy, efficiency, and robustness, providing theoretical guidance and a reference basis for future intelligent simulations in tunnel and underground engineering applications.

Shield tunneling in soft ground is often interrupted by equipment maintenance, ground improvement at shafts and portals, or unforeseen incidents, which can lead to stress redistribution and associated ground deformation. This paper quantifies the deformation response induced by shield shutdown and restart, with a focus on the superimposed disturbances during the two phases. The study is based on a shutdown-restart event on Tianjin Metro Line 7. A model for the shutdown stage was developed in ABAQUS and validated against field monitoring data. The validated framework was subsequently extended to replicate the complete sequence of face support pressure decay, shield sinking, and TBM restart, facilitating a systematic evaluation of deformation evolution in soft soils. The results indicate that during shutdown, the maximum settlement (Smax) increases to approximately five times its initial value. The extent of settlement influence expands from about 2D (where D is the tunnel diameter) to 5D, and the width of the settlement trough increases from approximately 2D to 4D. Shutdown-induced settlement increases with burial depth and decreases with distance from the excavation face, exhibiting a two-stage temporal pattern characterized by rapid early development followed by slower late-stage growth. During the restart phase, the secondary disturbance is primarily concentrated within the initial advance of 0–4 m; therefore, enhanced monitoring and control are recommended for the early restart segment and near-field sections around the face (approximately ±8 m). These findings provide a theoretical basis for parameter control and monitoring-warnings for shield shutdown-restart operations in soft soil.

The alteration of the seepage field is one of the most significant processes influenced by the construction of high dams, typically characterized by a wide range of changes, extensive areas of impact, and prolonged durations. Thus, characterizing the spatiotemporal evolution of the seepage field in dam and reservoir areas is crucial for the safety assessment of high dams. Based on seepage monitoring data collected over the past ten years at the Xiluodu super-high arch dam site, this study investigates the spatiotemporal distribution and variations of the seepage field and assesses the performance and safety of the seepage control system using numerical simulation and inverse modeling methods. The results indicate that the permeability of the near-dam fractured rock masses upstream of the grout curtains exhibits an exponential decay over time, having decreased to approximately 10% of its initial values and approaching a steady state. This permeability variation is primarily attributed to the clogging of suspended sediments transported by groundwater flow within the rock fracture networks. Consequently, the discharges from various parts of the site and the pore water pressure in the foundation rock masses upstream of the grout curtains decrease annually. Additionally, the hydraulic gradient of the grout curtains shows a noticeable decline, indicating a rising trend in seepage safety at the site rather than a decrease. The grout curtains and drains continue to perform effectively even after years of operation, ensuring that the seepage field is well-controlled. However, the discharge at the site is spatially unevenly distributed, influenced by factors such as geological structures, landform conditions, and unplugged boreholes drilled during site characterization. These findings are highly significant for the long-term seepage safety assessment of high dam projects.

To address the challenges associated with accurately identifying the integrity of tunnel surrounding rock in complex geological environments—specifically issues of high difficulty, insufficient refinement, poor timeliness, and low accuracy of advanced predictions—a three-dimensional refined intelligent identification and dynamic advanced prediction method for assessing the fragmentation degree of tunnel surrounding rock in mechanized drilling and blasting construction has been developed. This method integrates drilling parameters collected by the rock drilling trolley with a deep learning algorithm model, focusing on the construction process：(1) The research findings indicate a significant correlation between drilling parameters and rock mass integrity. Utilizing this relationship, the proposed method for extracting the fluctuation features of drilling parameter data can effectively characterize the development of joints and fissures within the rock mass. Unlike traditional surface observation methods at the tunnel face, this approach employs a spatial interpolation technique for discrete borehole data, enabling a non-uniform, high-density characterization of the surrounding rock’s integrity within the three-dimensional spatial range of the excavation face. (2) Taking into account the spatial distribution characteristics and temporal variation features of the drilling parameters, an intelligent recognition method for the integrity of tunnel surrounding rock using the drilling and blasting technique was constructed based on an attention spatial-temporal fusion network model, achieving an accuracy rate of 91.38%. This represents an improvement of 19.58% compared to the LSTM model, transforming the recognition of rock mass integrity at the tunnel face from a superficial understanding to a “three-dimensional refined dynamic intelligent evaluation.” (3) Based on the spatial continuity and gradual change characteristics of the lithology through which long tunnels pass, an advanced prediction model for the integrity of surrounding rock is proposed, utilizing a bidirectional long short-term memory neural network (Bi-LSTM). This model dynamically predicts the integrity of surrounding rock in the next unexcavated cycle based on the variation patterns of surrounding rock information from the excavated section, achieving an accuracy rate of 88.75%. It provides effective support for evaluating surrounding rock quality, identifying stability, and making dynamic decisions regarding excavation and support in mechanized tunnel construction using the drilling and blasting method.

Traditional methods for modeling discontinuity networks often struggle to accurately represent the discreteness and heterogeneity induced by weathering in rock masses. To address this issue, this paper proposes an enhanced stochastic discontinuity generation approach. The method integrates discontinuity data obtained from borehole tests and multi-view photogrammetry, quantitatively characterizes the degree of weathering using velocity ratios and a secondary calcite alteration index, and establishes a functional relationship between the degree of weathering and depth. Subsequently, a heterogeneous discontinuity distribution model is developed to improve the generation process of discontinuity center points, allowing for variations in discontinuity distribution characteristics across different weathering degrees and enabling the simulation of heterogeneous discontinuity center points. Comparative results with conventional simulation methods demonstrate that the proposed approach can accurately reproduce the RQD values of rock masses, with the simulated discontinuity volume density and connectivity characteristics showing a reasonable alignment with the discontinuity distribution characteristics of natural rock masses. The proposed method is capable of simulating discontinuity distribution characteristics under varying weathering conditions and can be further applied to numerical simulations and stability evaluations, offering new theoretical support for the engineering stability analysis of complex weathered rock masses.

To address the low drainage efficiency and insufficient consolidation typically encountered during the vacuum preloading of low-permeability soft soil, this study introduces a pre-vacuum freeze-thaw disturbance aimed at enhancing the soil structure and improving drainage performance. Vacuum drainage tests were conducted on specimens subjected to various disturbance modes, with continuous monitoring of the coupled thermal-hydraulic-mechanical responses, settlement evolution, and energy consumption throughout the artificial freeze-thaw and vacuum stages. Post-consolidation measurements of water content and shear strength were also performed. The results indicate that ice crystallization and fissure development during freezing significantly increased soil permeability by approximately 10%-40%. Upon thawing, the newly formed pore channels facilitated accelerated vacuum drainage. During vacuum preloading, the freeze-thaw disturbed specimens exhibited faster water content reduction, more rapid dissipation of pore-water pressure, greater strength enhancement, and larger settlement compared to the undisturbed specimens, indicating a significant improvement in overall consolidation efficiency. Although the disturbed group consumed more total energy, its energy-consumption ratio throughout the process displayed a stage-dependent pattern: an initial sharp rise, a rapid mid-stage decline, and a slight late-stage increase. Notably, the late-stage energy consumption ratio of the disturbed specimens was lower than that of the undisturbed group, demonstrating superior full-cycle energy performance. This study provides a novel technical pathway for achieving efficient dewatering and consolidation of soft soil.

This paper presents a comparative analysis of the evolution of elastoplastic characteristics (including resilient modulus, MR and accumulative plastic strain, ) of a clayey subgrade soil in Guangxi Ningming under two cyclic loading-wetting paths. The applied loading-wetting paths include a “loading-wetting” (LW) path, which simulates the simultaneous cyclic loading and wetting processes during the service of pavement subgrade soils, and a “constant-humidity loading” (CH) path, which is commonly used in current tests where soils are loaded under constant moisture content conditions. The experimental results indicate that: (1) compared with the CH path, the clayey soil under the LW path exhibits stronger resistance to cyclic loading, demonstrated by a higher MR (up to 39% increase) and a lower (up to 57% reduction) at the same moisture content and stress state. Moreover, under the LW path, the MR increases more rapidly during cyclic loading, while the accumulates more slowly. The scales of variation in the MR and during the wetting process are also smaller than those under the CH path; (2) Under both paths, the MR and generally exhibit a linear relationship during cyclic loading, with the slope decreasing significantly as the moisture content (w) increases. At lower w values, the accumulation of the and the corresponding compaction effect effectively enhance the MR. At higher w values, the accumulates rapidly, but the cyclic loading-induced damage effect increases while the contribution of the compaction effect decreases, resulting in insignificant MR growth. The slope of the MR-  relationship under the LW path is greater than that under the CH path; (3) At the same number of loading cycles, the MR-  relationship of the clayey soil under different confining pressures, deviatoric stresses, moisture contents, and loading-wetting paths shows a distinct linear trend in double-logarithmic coordinates. Based on this, a model is proposed that effectively predicts the MR-  relationship of clayey subgrade soils under varying numbers of loading cycles, moisture contents, and stress states.

This study investigates the behavior of in-situ weak expansive soil as a road subgrade foundation under coupled hydrological and mechanical actions. A series of drying-wetting cycle tests were conducted under varying overburden pressures and moisture variation amplitudes. The results reveal the coupled effects of load and moisture amplitude on the volumetric response of the foundation soil. Subsequently, isotropic consolidation and true triaxial shear tests were performed under constant water content using an unsaturated soil true triaxial apparatus to elucidate the regulatory mechanisms of load, drying-wetting history, and intermediate principal stress on the strength characteristics of the in-situ soil. The key findings are as follows: the coupled actions of load and moisture amplitude dictate the direction of cumulative deformation. Specimens exhibit cumulative expansion under zero load, with the magnitude increasing with larger moisture amplitudes. In contrast, specimens transition to cumulative contraction when a load is applied, and this contraction intensifies with higher load or moisture amplitude. The load alters the evolution pattern of swell-shrink magnitude and exerts a continuous suppressive effect on volumetric changes. The critical water content (wc) increases with higher load, moisture amplitude, and the number of cycles. The wc-N curves under different loads form hysteresis loops, indicating an irreversible path dependency of the swell-shrink characteristics on load history. The peak strength degrades with an increasing number of drying-wetting cycles but improves with higher load, intermediate principal stress, or smaller moisture amplitude. All three factors effectively decelerate the rate of strength deterioration. Strength stabilization occurs prior to deformation stabilization. Particularly under conditions of low load or small moisture amplitude cycles, the long-term cumulative deformation risk must be assessed independently. The research outcomes provide a theoretical basis for foundation treatment design and long-term performance prediction of subgrades in expansive soil regions.

To investigate the evolutionary characteristics of earth pressure and the synergistic deformation control mechanism of inclined pile-supported excavations during foundation pit excavation, eight large-scale model tests were conducted. The results demonstrate the following: (1) The inclined alternating pile system reconstructs the load transfer path through a spatial truss effect, significantly reducing pile head displacement, which decreases further as the pile inclination angle increases; (2) The inclination of piles induces a redistribution of the principal stress field, wherein the active earth pressure approaches the Rankine theoretical value, while the passive zone develops a new trapezoidal pressure pattern due to progressive failure; (3) The three-dimensional interaction of the inclined alternating piles reduces the measured passive earth pressure to 34.42%-57.82% of the Rankine value. Additionally, the asymmetric stiffness of the structure delays the development of the plastic zone; (4) As excavation depth increases, earth pressure evolves from elastic adjustment to limit equilibrium. Under combined compression and bending, the piles exhibit elastic bending responses, and the deformation of the soil behind the wall is non-uniformly distributed, leading to larger deformations required to reach active and passive limit states compared to rigid retaining walls. These findings reveal the evolution mechanism of earth pressure and the deformation regulation effect of inclined pile-supported systems, providing experimental evidence and a theoretical reference for optimizing the design and safety evaluation of deep excavation support structures with inclined piles.

To enhance the design methodology for reinforced soil retaining walls, this study proposes a performance-based design framework and methodology for rigid facing reinforced soil retaining walls, guided by the functional requirements of railway subgrades and informed by the three-level seismic fortification philosophy employed in building seismic design. This framework aims to meet the millimeter-scale deformation control requirements of high-speed railways. The findings reveal that: (1) a multi-level performance requirement system encompassing safety, serviceability, and repairability has been established, with clearly defined performance indicators and verification items. Performance verification methods aligned with these requirements have been developed, and strength-based control criteria have been introduced to ensure collapse prevention during rare earthquakes. (2) Graded control limits for structural settlement, which are directly related to train operating speed, have been established. A method for calculating the strain and tensile force of individual reinforcement layers based on facing deformation has been developed, leading to a coordinated multi-parameter verification mechanism centered on surface settlement that integrates facing internal forces and reinforcement behavior. This approach marks a fundamental shift in the design of reinforced soil retaining walls from a singular focus on strength-based control to a coordinated approach addressing multiple performance objectives. (3) The design tensile strength of reinforcement specified in the Chinese code is significantly lower than that in the Japanese code, resulting in more stringent requirements for reinforcement length and a conservative design approach. Calculation examples demonstrate that the Chinese method necessitates a 50% increase in reinforcement length and a 62.5% increase in reinforcement strength. It is recommended that the Japanese design methodology and relevant correlation coefficients be considered when determining the design tensile strength of reinforcement to enhance design economy. (4) Backfill cohesion is a key parameter influencing structural stability. Calculations indicate that increasing the cohesion from 0 to 20 kPa can enhance the sliding and overturning stability coefficients (K) by approximately 200%. It is advised that this beneficial effect be appropriately considered in practical design to achieve a balance between safety and economy. This study offers systematic theoretical and methodological support for the safe implementation of rigid facing reinforced soil retaining walls in high-seismic-intensity regions of China for high-speed railways.

The performance evaluation of clay liners for seepage prevention and isolation is widely applied in engineering fields, including municipal solid waste landfills, hazardous waste landfills, and remediation of contaminated sites. However, most existing studies have only conducted simple coupling analyses based on 2 to 3 physical processes among the Thermal-Hydro-Mechanical-Chemical (THMC) multi-physics framework, often neglecting the nonlinear variations of several important parameters in their models. To investigate the mutual feedback coupling mechanisms of the THMC multi-physics involved in clay liners, this study adopted a large-strain nonlinear consolidation model. This model incorporates the effects of chemical osmotic pressure induced by contaminant transport on soil consolidation, as well as the impacts of temperature rise on the physical-mechanical properties of soils and contaminant transport characteristics. A comprehensive large-strain nonlinear model for heat transfer, consolidation, and contaminant transport was established within the context of THMC multi-physics mutual coupling, along with its numerical solution. The correctness and applicability of the proposed coupling model were verified by comparing contaminant transport characteristics and soil consolidation settlement under specific scenarios. Based on this analysis, the mutual feedback coupling mechanisms and influencing factors among the THMC multi-physics in clay liners were thoroughly discussed. The results indicate that the introduction of coupling effects from temperature and chemical interactions on consolidation and contaminant transport accelerates soil settlement rates, increases peak settlement, and exacerbates the “over-consolidation” phenomenon caused by chemical loads. Additionally, the incorporation of thermal dynamics not only accelerates the dissipation of excess pore water pressure but also enhances the contaminant transport rate, with the effects becoming more pronounced over time. Furthermore, chemical osmotic pressure can induce negative excess pore water pressure, resulting in the overall excess pore water pressure of the soil remaining negative even during the long-term consolidation stage, which complicates complete dissipation post-consolidation. Finally, the model was applied to an unregulated landfill, demonstrating that the established model possesses strong engineering applicability.