This study addresses the challenge of disposing of mud waste generated by underground engineering projects in soft-soil regions. A solid alkali-activated geopolymer based on slag and fly ash was used to stabilize mud with an initial moisture content of 80%, thereby converting it into recycled fine aggregates. First, at the macroscopic scale, this study evaluated the effects of key geopolymer mixture parameters—including precursor proportion, precursor content, the molar ratio of solid sodium silicate, alkali-activator content, and the water-to-precursor ratio—on the pore distribution, connectivity, and particle-packing behavior of the recycled fine aggregates. These findings provide a theoretical basis for their lightweight engineering application of these aggregates. Subsequently, drop-hammer impact tests were conducted to investigate how these parameters affect the impact-breakage behavior of the aggregate fills. The analysis focused on post-impact changes in particle-size distribution, fractal dimension, and relative breakage index. At the microscopic level, scanning electron microscopy and energy-dispersive X-ray spectroscopy were used to characterize the micromorphology and chemical composition of individual recycled fine-aggregate particles. These observations clarified the mechanisms governing the impact-breakage behavior of recycled fine aggregates with different mixture designs and established a cross-scale correlation between macroscopic performance and microstructure. The results show that although all trial fills satisfy the lightweight requirements for protective cushion layers (bulk density <1 200 kg/m3), their basic physical properties and impact-breakage behavior differ significantly. These differences are governed primarily by the type of geopolymer gel formed and by the morphology and spatial distribution of the pores. Notably, when the precursor contains 90% slag and 10% fly ash at a dosage of 30%, together with solid sodium silicate at a molar ratio of 0.8 and an alkali-activator content of 15.85%, direct dry mixing with mud at an initial moisture content of 80% yields recycled fine aggregates with excellent physical properties and outstanding resistance to impact breakage. The findings of this study identify a promising granular material for the protective cushion layers of underground structures. They also promote the high-value utilization of mud resources and demonstrate significant engineering and environmental benefits.

Artificial ground freezing (AGF) is widely applied in geotechnical engineering under complex site conditions. However, in situ methods for evaluating the strength of frozen soil are still insufficiently developed. This study employs a self-developed embedded four-electrode system to investigate the strength and electrical properties of water-rich silty clay collected from the Yangtze River floodplain. Results show that unconfined compressive strength (UCS) increases as temperature decreases, following a “slow rise–sharp increase–slow rise” pattern, and also increases with dry density. A critical water content of 22% produces the peak strength. Beyond this threshold, frost-heave-induced cracking reduces the strength. Under near-saturated conditions, the formation of an ice-skeleton structure partially restores the strength. Resistivity increases in stages as temperature decreases, primarily because of free-water freezing and increased pore tortuosity. At higher water contents, conductivity diminishes more sharply. A transition in the conduction mechanism occurs at a critical dry density, with conduction shifting from pore-water dominance to bound-water and soil-matrix dominance. A power-law relationship is observed between saturation and resistivity, and the freezing of bound water together with frost deformation further enhances this sensitivity. Unconfined compressive strength and deformation modulus are strongly linearly correlated, and both increase with resistivity. A resistivity-based predictive model is established to link macroscopic strength to microstructural evolution and moisture variation, thereby explaining the model dispersion and predictive accuracy from a microscopic perspective. Brunnauer-Emmett-Teller (BET) and Scanning electron microscope (SEM) analyses reveal that increasing water content enlarges the pore structure. In addition, freeze–thaw cycles transform soil aggregates from flaky to honeycomb-like forms and promote the development of micropores into continuous macropores. These microstructural changes illustrate the damage process and demonstrate how the coupling between water–ice phase transitions and pore evolution influences soil strength. This study enhances understanding of the multi-parameter coupling between frozen soil properties and resistivity, thereby supporting the development of accurate resistivity-based evaluation models for artificially frozen ground.

Environmental problems associated with the large volumes of waste soil generated by large-diameter slurry shield tunneling have become increasingly severe. MgO-based solidification-carbonation is an efficient, environmentally friendly, and sustainable method for modifying waste soil. However, the governing patterns and mechanisms of mass transfer during carbonation remain unclear. To address this issue, this study quantitatively determined the pore water and CO2 contents in carbonated soils with different MgO contents and carbonation times based on the water transformation mechanism. It then investigated the mass-change behavior of each soil component, proposed a mass transfer mechanism and a mass transfer model, and analyzed the relationships between the model and the mechanical properties of the soil. The results show that soil mass changed during carbonation consist of CO2 absorption and pore water evaporation. The variations in the contents of each soil component conforms to the water transformation mechanism and the synergistic carbon–water interaction. The proportion of soil mass increase caused by CO2 absorption increased with the prolongation of carbonation time, with a maximum absorption capacity of 197 kg/m3. Meanwhile, water transformation gradually shifted toward the formation of mineral water, with up to 48% of pore water converted into carbonation-derived mineral water. Fitting results indicated that, as the MgO content increased, the ratio of pore water consumption to CO2 absorption decreased, suggesting improved CO2 utilization efficiency during carbonation. By establishing a mass transfer model based on the water transformation principle, the changes in unconfined compressive strength (UCS) and elastic modulus could be indirectly inferred from variations in the individual components of the model. Among these variables, pore water variation exhibited an exponential relationship with UCS and a linear relationship with elastic modulus. The calculation method for analyzing mechanical properties through the mass transfer model provided a novel approach for predicting the properties of MgO-based solidified/carbonated high-plasticity silt generated by slurry shield tunneling.

Soil heavy metal (HM) contamination is a widespread and serious environmental challenge. Even after treatment, contaminated soil may still pose a risk of secondary metal release under complex environmental conditions. In this study, a novel solidifying agent, a diatomaceous earth–fly ash-based geopolymer (D-FA), was synthesized from phosphoric acid (H₃PO₄), fly ash, and diatomaceous earth for the solidification/stabilization of contaminated lead-zinc red clay. Under coupled acid exposure and wet–dry cycling conditions, short-term effectiveness was evaluated using the toxicity characterization leaching procedure (TCLP), whereas the accelerated TCLP (ATCLP) was applied to analyze leaching kinetics and predict long-term performance. Scanning electron microscopy (SEM), energy-dispersive spectroscopy (EDS), and X-ray diffraction (XRD) were employed to characterize microstructural evolution and degradation mechanisms across multiple scales. A durability assessment framework based on cumulative leaching fraction (CFL) and effective diffusion coefficient (Dₑ) was established to predict the environmental safety period. The results showed that D-FA increased the strength of red clay by 6.59-fold and achieved fixation efficiencies of 99.99% for Pb2+ and 73.95% for Zn²⁺ within 28 days. Under long-term acid exposure and wet–dry cycling conditions, D-FA exhibited greater stability and durability in immobilizing Zn²⁺ than Pb²⁺. According to predictions based on Fick’s diffusion model, the environmental safety period reaches several decades for farmland at pH=5, exceeds 200 years for grassland and forest land, and may extend to several centuries for construction sites.

This study systematically investigates the swelling behavior of red-bed mudstone from the Jurassic in Central Sichuan, under overlying loads using macroscopic water-absorption expansion tests, microstructural observations, and molecular dynamics simulations. The results demonstrate that the expansion process of red-bed mudstone can be divided into three distinct stages: rapid, decelerated, and stable. With increasing overlying load, both the macroscopic expansion force and water absorption rate generally exhibit a linearly decreasing trend; however, an anomalous peak appears at 100 kPa. This anomaly is attributed to critical loading-induced microcracking, which facilitates water penetration and enhances mineral hydration. Microstructural analysis indicates that within the micropore range, the fractal dimension of the pores reaches a maximum value at 1.88 under 100 kPa. This finding reflects a positive correlation between microstructural complexity and macroscopic expansion. Molecular dynamics simulations further show that, at the nanoscale, applied loading inhibits the diffusion of interlayer water molecules and ions in clay minerals. The diffusion coefficients of water molecules and Na⁺ decreased to 3.3×10⁻⁷ cm²/s and 2.0×10⁻⁷ cm²/s, respectively. This inhibition enhances the stability of the hydrated structure. This study elucidates the internal mechanism and evolutionary patterns of water-absorption expansion in red-bed mudstone across macro–micro–nano scales. It provides a theoretical basis for geological hazard prevention and engineering stability assessment in red-bed strata.

The stability of deeply excavated canal slopes in expansive soil is governed by groundwater-level fluctuations, wetting–drying cycles, and inherent fissures. To investigate these effects, a hydraulic response model test was conducted on a slope with inherent fissures. The left side of the slope contained fissures, whereas the right side remained intact for comparison. The experiment included four wetting–drying cycles, groundwater-level variations, and one prolonged rainfall event. The results showed that: 1) fissures altered the water distribution, resulting in pronounced spatial variations in water content on the left side, whereas the right side became more uniform after prolonged rainfall; 2) the pore water pressure in the deep soil was primarily controlled by the groundwater level, and the response was stronger on the left side; 3) the matric suction of the deep soil on the left side exhibited a response pattern different from that induced by wetting–drying cycles alone; 4) matric suction continuously decreased with increasing wetting–drying cycles, reducing shear strength and accelerating slope displacement, with a lag in the displacement response. The combined effects of groundwater-level fluctuations and inherent fissures intensified the deformation and failure of the slope.

Deep oil and gas resources have considerable development potential. However, reservoir fracturing is affected by high confining pressure. The fracture tip exhibits pronounced nonlinear mechanical behavior, which hinders hydraulic fracture propagation. To clarify the nonlinear fracture characteristics of rocks under confining pressure, three-point bending fracture experiments were conducted on standard single-edge notched beam specimens of yellow sandstone. The evolution of the fracture process zone (FPZ) was monitored using an acoustic emission system. The results show that rock exhibits significant nonlinear fracture characteristics under confining pressure, and that higher confining pressure prolongs the entire pre-peak evolution of the FPZ. At high confining pressures, the microscopic fracture mechanism in rock shifts from intergranular to transgranular fracture. This transition in the microscopic fracture mode is the primary reason for the smoother fracture surface and the increase in fracture energy. Based on the cohesive zone model (CZM), an analytical model was established to predict the FPZ length. As confining pressure increases, the FPZ length decreases, whereas the fracture energy increases. This indicates that confining pressure not only limits the extent of FPZ development but also significantly enhances the energy dissipation intensity per unit area within the zone. Macroscopically, this manifests as a requirement for a higher stress level to drive further FPZ development and fracture propagation. This phenomenon reflects the strengthening effect of confining pressure on the nonlinear fracturing behavior of rock. The research results provide theoretical support for stimulation and modification of deep reservoirs.

Understanding the mechanical anisotropy and damage evolution of layered sandstone under cyclic loading is crucial for evaluating the dynamic stability of underground rock engineering. In this study, seven rectangular sandstone specimens with different bedding dip angles were subjected to staged cyclic loading tests. By integrating strain monitoring in different orientations, energy decomposition, and acoustic emission signal monitoring techniques, the mechanical anisotropy, energy evolution, and acoustic emission characteristics of layered sandstone under staged cyclic loading were analyzed. The results show that: 1) The bedding dip angle significantly affects the anisotropic response characteristics of sandstone strength and deformation. At low and high bedding dip angles, the sandstone exhibits relatively high strength, a low strain anisotropy coefficient (g), and predominantly tensile failure. In the intermediate dip angle range (30º, 45º, 60º, 75º), g increases, and shear failure becomes dominant. For a given bedding dip angle, the longitudinal strain anisotropy coefficient (g_A) is greater than the transverse strain anisotropy coefficient (g_L). 2) The total input energy and elastic energy calculated from the longitudinal strain on the strike bedding plane (e_SA) show a U-shaped relationship with bedding dip angle, whereas those calculated from the longitudinal strain on the dip bedding plane (e_DA) exhibit a Λ-shaped relationship. This finding indicates that energy release is closely associated with bedding-plane orientation. 3) The damage variable constructed based on cumulative ringing counts reveals remarkable anisotropy in damage evolution. The rise-time/amplitude ratio (RA) and average frequency AF (RA-AF combined analysis) can accurately distinguish crack types. The proportion of shear cracks exhibits a Λ-shaped distribution, peaking at a dip angle of 75º, which corresponds to the U-shaped distribution of macroscopic strength. Acoustic emission location technology can effectively characterize the spatial distribution of final damage. Therefore, this study investigates the effects of bedding dip angle on sandstone mechanical anisotropy, bidirectional energy distribution, and damage evolution characterized by multiple acoustic emission parameters. The results are expected to provide theoretical support for dynamic stability evaluation and disaster early warning of layered surrounding rock.

Breakthrough pressure serves as a critical parameter for characterizing the capillary sealing capacity of caprock formations. The injection of CO₂ into subsurface reservoirs triggers complex interactions among CO₂, formation brine and rock. These interactions have been shown to significantly affect the rock’s transport properties and sealing performance. However, the mechanisms by which CO₂-brine-caprock interactions affect caprock breakthrough pressure remain poorly understood. To elucidate the response mechanism of caprock breakthrough pressure to CO₂–brine–rock interactions, breakthrough pressure experiments were performed over different durations of CO₂–brine–rock interaction, utilizing two distinct types of caprock cores obtained from the Junggar Basin. A suite of complementary analytical methods, including nuclear magnetic resonance, scanning electron microscopy, liquid chromatography, and contact angle measurements, was employed to characterize the evolution of pore structure and wettability within the cores. The results show that CO₂-brine-rock interactions reduce the breakthrough pressure of the caprock samples. This reduction is attributed to the combined effects of increased pore volume, enhanced permeability, and diminished wettability. Further kinetic analysis of mineral dissolution suggests that the decline in breakthrough pressure slows over time and may eventually reach a steady state. This behavior is likely attributable to the inhibitory effect of secondary mineral precipitation. These finding underscore the critical importance of accounting for CO₂-brine-rock-induced degradation of caprock sealing capacity when defining pressure thresholds for the safe operation of CO₂ geological storage projects, thereby helping to ensure long-term stability and safety.

Current theoretical studies on the trench stability of diaphragm walls seldom consider the coupled action of rainfall and water table fluctuations. Quantitative understanding of the evolution of soil mechanical properties under intermittent rainfall infiltration remains limited. First, on the basis of the Green-Ampt model combined with the Bodman-Colman layered hypothesis, a modified Green-Ampt infiltration framework is developed to simultaneously capture the degradation of saturated hydraulic conductivity caused by wetting-drying cycles and evaporative losses during rainfall interruptions. In this framework, a fractal pore-structure analysis is employed to derive a predictive equation for the saturated hydraulic conductivity of unsaturated soils. This equation, in turn, governs the dynamic propagation of the wetting front. The Penman-Wilson equation is then used to quantify evaporation during hiatus periods and to update the initial volumetric water content in the transition zone. Second, by coupling Darcy’s law with the Mohr-Coulomb criterion, a piecewise formulation for apparent cohesion is established as a function of rainfall intensity and duration. Third, an upper-bound limit-analysis procedure is developed that incorporates a three-dimensional rotational (log-spiral) failure mechanism derived from a spatial discretization scheme. The mechanical work associated with apparent cohesion is introduced into the virtual-power equation, yielding an explicit upper-bound factor of safety for trench walls subjected to intermittent rainfall infiltration and water-table oscillations. Finally, comparisons with numerical simulations, previously reported full-scale experiments, and field monitoring data show excellent agreement, thereby verifying the predictive capability of the proposed model. In addition, a sensitivity analysis was conducted on key intermittent-rainfall parameters, including rainfall intensity, number of rainfall events, and average temperature, as well as key soil and geometric parameters relevant to diaphragm-wall stability, including pore-size distribution characteristics, wall aspect ratio, effective cohesion, effective internal friction angle, and groundwater depth. Parametric studies reveal that higher rainfall intensity, more frequent rainfall events, and elevated ambient temperature deepen the wetting front and reduce the factor of safety. A shallower water table markedly decreases stability, whereas a sufficiently high slurry level provides partial support. A larger pore-distribution parameter l, a smaller fractal dimension D, and a trench width-to-depth ratio of L/H<1.5 exacerbate 3D edge effects and reduce stability. Conversely, increases in effective cohesion and effective friction angle markedly enhance shear resistance, and edge effects become negligible when L/H>10. These findings provide a rigorous theoretical basis and practical guidance for risk assessment and support-parameter optimization in diaphragm-wall construction under coupled intermittent rainfall and water-table fluctuations.

Anisotropic consolidation significantly affects the liquefaction behavior of saturated sand. To investigate the evolution characteristics of anisotropically consolidated saturated coral sand under cyclic loading, undrained cyclic shear tests were conducted with a 90º jump rotation of the principal stresses under varying consolidation conditions. The test results indicate that consolidation conditions significantly affect the stress-strain response of saturated coral sand. The hysteresis curve evolves with unidirectional accumulation when the stress component is collinear with the consolidation stress, whereas alternating cyclic hysteresis responses are observed under misaligned stress orientations. Based on variations in the average flow coefficient and its gradient with the number of cycles, the mechanical evolution of coral sand can be classified into three phases. The initial phase corresponds to an elastic solid state, characterized by low excess pore water pressure and negligible mobility. This is followed by a phase of accelerated pore water pressure accumulation, leading to a solid-liquid transition with increased particle mobility. Finally, a sharp change in the average flow coefficient gradient signifies the onset of a transient viscous-liquid state. Under varying consolidation and cyclic-loading conditions, gradient shows an initial increase followed by a decrease with increasing excess pore water pressure ratio with a turning point at a critical excess pore water pressure ratio r_uth of approximately 0.9, which remains invariant across test conditions. Given the observed discrepancy between pore water pressure development and strain response in anisotropically consolidated coral sand, this r_uth may serve as a reliable triggering criterion for liquefaction in saturated sandy soils under complex static and dynamic loading.

Compacted bentonite is widely used as a hydraulic barrier in municipal landfills and deep geological repositories for nuclear waste. However, long-term exposure to salt-bearing pore fluids can substantially increase its permeability, thereby threatening the barrier integrity and containment safety. Therefore, developing a theoretical model that quantitatively describes how permeability depends on salt concentration and mechanical loading is critical for performance prediction. In this study, we investigate the microstructure of bentonite fully saturated with saline solutions, focusing on the alteration of interlayer and inter-aggregate pores induced by variations in ionic strength. We then revise the conventional relationship between permeability and montmorillonite void ratio by explicitly considering the salt effects on the diffuse double layer and interparticle forces. Additionally, a modified effective stress is introduced to account for physicochemical interactions, and a fractal model is formulated to capture the coupled influences of overburden pressure and solution concentration on the permeability. The fractal dimension in the model reflects the tortuosity and connectivity of the pore network, both of which are sensitive to salt concentration. The model is validated using experimental data from our consolidation-permeability tests and from previously reported studies, covering a range of NaCl concentrations and effective stress levels. The comparison demonstrates that the relative errors between the predicted and measured values are within ±12%, and the overall agreement is highly satisfactory. The proposed model provides a simple, reliable, and physically grounded tool for estimating the evolution of bentonite permeability under chemically aggressive conditions, thereby facilitating long-term safety assessment and engineering design.

Compared with traditional blasting methods, static cracking is safer and more environmentally friendly and has been widely used in stone quarrying, rock fracturing, and urban building demolition. However, research on its directional fracture performance remain limited. Challenges such as long crack initiation times and poor predictability and controllability of fracture propagation direction persist. To investigate the influence of notch depth on the fracture behavior of specimens under expansive loading, five groups of static cracking tests were conducted on concrete specimens with different notch depths. Acoustic emission (AE), strain monitoring, and digital image correlation (DIC) were combined to examine the effects of notch depth on fracture modes under expansive loading. The evolution of energy characteristics, strain characteristics, and surface strain fields at different failure stages was also analyzed. Two types of cracks were observed during the tests: P-type cracks, which propagated parallel to the slot direction, and T-type cracks, which propagated perpendicular to it. The results show that, compared with circular-hole specimens, double-wing slotted-hole specimens exhibited crack initiation approximately 1 h earlier. In addition, greater notch depths led to earlier crack initiation. As notch depth increased, the final fracture mode gradually shifted from P-P-T-T to P-P, indicating that deeper notches enhanced stress concentration at the slot and thereby suppressed the formation of T-type cracks. In the P-P-T-T and P-P-T fracture modes, the initiation of T-type cracks caused the strain on the corresponding free surface to change from an increasing to a decreasing trend, producing a distinct turning point in the stress state. When P-type cracks initiated, both the AE hit count and cumulative energy increased sharply. Moreover, the AE hit count and cumulative energy were higher in the P-P-T-T and P-P-T fracture modes than in the P-P mode. In addition, the initiation of T-type cracks produced a second peak in the AE hit count.

With the widespread construction of high dams and large reservoirs, together with increasingly complex operating environments, dam safety early warning systems are facing greater demands. There is an urgent need to address the limitations of traditional methods, which often lack predictive capability under complex conditions. This paper systematically reviews recent advances in intelligent dam safety early warning, with a focus on three core aspects: intelligent risk assessment, the intelligent development of early warning indicators, and risk warning models based on intelligent optimization. The study examines the evolution of risk assessment from static threshold-based judgment to multidimensional probabilistic state characterization. It also describes the shift in early warning indicator selection from manual, experience-based screening to data-driven automatic optimization. Furthermore, it highlights the critical role of intelligent optimization algorithms in addressing parameter sensitivity and local optimum problems in machine learning and deep learning models. Research indicates that integrating intelligent optimization strategies can significantly reduce reliance on empirical parameters while improving model generalizability and stability in complex nonlinear environments. This study aims to advance dam safety early warning systems toward real-time, accurate, and adaptive operation, thereby providing a solid theoretical foundation and key technical support for the development of a new generation of intelligent dam safety warning frameworks.

The mechanism governing the transition from creep to rapid movement is a central challenge in understanding the instability of reservoir landslides. This study focuses on the Muyubao landslide, a giant bedding landslide in the Three Gorges Reservoir Area that exhibits persistent creep. This study focuses on the Muyubao landslide, a giant bedding landslide exhibiting persistent creep in the Three Gorges Reservoir Area. A series of ring-shear tests were conducted on the shear-zone soil, revealing that its residual friction coefficient depends on both normal stress and displacement rate. Specifically, friction weakening, characterized by exponential decay, occurred at low displacement rates, whereas slight friction strengthening, characterized by logarithmic increase, emerged when the displacement rate exceeded a critical threshold (v = 3.33 × 10⁻⁴ m/s). The research indicate that friction weakening is the dominant mechanism driving the transition from slow creep to catastrophic failure by promoting accelerated creep through a positive feedback loop of “rate increase–friction weakening–further rate increase.” Based on the experimental results, a 3D slice-based dynamic calculation model was established that incorporates normal stress, rate-dependent frictional behavior, and variations in pore water pressure. This study reveals the potential deformation and instability mechanisms of the MYB landslide and provides essential theoretical support for predicting instability velocity and evaluating the movement tendency of creep-type landslides in reservoir areas.

Under the influence of the super-thick nappe overlying the coal seam, the fracture morphology of the overburden and the stress distribution change significantly after the extraction of the working face. The abnormal concentration of stress leads to large deformations in the roadway and increases the risk of roof collapse, thereby restricting the safe extraction of the working face. This paper takes the mining of 360801 working face under Fufeng super-thick nappe in Xinji No.1 Coal Mine as the engineering background. Firstly, similar material tests were conducted to investigate the structural evolution of the overlying strata under the influence of the super-thick nappe. It was found that after mining, the super-thick nappe above the working face bent and subsided as a plate-like structure. Meanwhile, the underlying strata developed an arching effect, forming an asymmetric pressure arch. Then, numerical simulation was used to analyze the stress distribution in the surrounding rock of the working face under the influence of the super-thick nappe. The simulations indicated that the asymmetric pressure arch in the overlying strata led to a markedly asymmetric stress distribution in the working face. The peak stress zone continuously shifted towards the front of the roadway, and the stress concentration increased as the face advanced. Compared with the non-nappe condition, the maximum displacements of the two sides and of the roof and floor in the roadway-intensive area increased by 32% and 52%, respectively. In addition, the peak stress increased by 8.86 MPa, and its influence distance ahead of the working face increased by 80 m. Finally, a mechanical model of the roof asymmetric pressure arch under the super-thick nappe was established, and the trajectory equation of the arch was derived. The influence range of the pressure arch was obtained, showing that with the increase of lateral horizontal stress, coal seam thickness, and burial depth, the range of pressure arch influence gradually expands. Furthermore, the reasonable stop-line for mining was determined to be 240 m from the roadway dense area. Field measurements showed that the maximum loads on the anchor cables and bolts supporting the surrounding rock in the roadway-intensive area were 176 kN and 38 kN, respectively. The maximum convergences of the roof-to-floor and side-to-side were 198 mm and 118 mm, respectively, representing reductions of 43.4% and 26.9% compared with those observed during final mining of the adjacent 360803 working face. The stability of the surrounding rock was significantly improved. The study provides a theoretical basis for designing stop-line positions and controlling surrounding rock in working faces under similar geological conditions.

Submarine landslides pose a severe threat to the safety of seabed infrastructure and coastal cities. Compared with subaerial landslides, submarine counterparts in deep water commonly occur on gentle slopes with gradients less than 5º and involve much larger sliding mass volumes. The progressive failure of these large-scale landslides is difficult to reproduce using traditional methods such as limit equilibrium and strength reduction methods. By incorporating the strain-softening behavior of cohesive soils into theoretical analyses and numerical simulations, the progressive failure of gentle slopes driven by shear band propagation can be effectively captured. Consequently, this framework has attracted growing attention in recent years. This paper systematically reviews recent advances in shear band propagation methods and summarizes the principles and applicability of key theoretical frameworks, including linear elastic fracture mechanics, energy-balance methods, and process-zone approaches. The review also discusses analytical and numerical solutions for complex conditions, including dynamic effects, two-dimensional curved slopes, three-dimensional slopes, deep-seated slip surfaces, and presents quantitative criteria for slope instability. The potential of shear band propagation methods for practical applications to submarine landslides is highlighted with reference to typical triggering factors such as rapid sedimentation, toe erosion, gas migration, and earthquakes.

Completely decomposed granite exhibits strong cohesiveness because of its high fine-particle content, posing a major challenge to earth pressure balance shield tunneling in metro projects in the southeastern coastal region of China. This material tends to adhere to the cutterhead and form mud cakes, which can sharply increase cutterhead torque, hinder muck discharge, and reduce excavation efficiency. In severe cases, frictional heating may solidify the mud cake, forcing shutdowns for chamber opening and creating substantial economic and safety risks. Foam conditioning is an effective approach for treating such cohesive strata. However, existing studies lack a robust strategy for systematically optimizing foam components for completely decomposed granite strata. As a result, it is difficult to balance multiple objectives, including foaming performance, conditioning effectiveness, and cost-effectiveness. In addition, the complex interaction effects within multicomponent foam systems have often been overlooked. To address this issue, this study employed response surface methodology (RSM) combined with central composite design (CCD) to develop a predictive model for foam performance. The independent variables were the anionic surfactant sodium alpha-olefin sulfonate (AOS), the nonionic surfactant alkyl polyglucoside (APG), the foam stabilizer xanthan gum (XHG), and the inorganic dispersant sodium hexametaphosphate (SHMP). In terms of performance evaluation, foaming ability was characterized by the foam volume measured with a modified Ross–Miles foam tester. Foam stability was assessed using both foam half-life and the 5-min defoaming rate. The ability to reduce soil adhesion was quantified by the interfacial adhesion force measured with a modified Yida conical bao pull-off tester. Multi-objective decision-making was performed using the desirability function approach, and the optimal compound formulation was identified as the Fuyan-1S foaming agent (FY-1S), containing mass percentage contents of AOS, APG, XHG and SHMP are 4.18%, 1.59%, 0.16% and 0.95%, respectively. Laboratory tests showed that the FY-1S foaming agent outperformed three selected commercial foaming agents in key properties, particularly stability and reduction of interfacial adhesion. Field engineering validation was conducted in an EPB shield tunneling section of Xiamen Metro Line 3. By comparing the conditioning effects of FY-1S with those of commercial foaming agents, this study systematically evaluated both soil properties, including interfacial adhesion force, undrained shear strength, and vertical slump, and shield tunneling parameters, including cutterhead torque, total thrust, advance rate, and chamber pressure. The results demonstrated that FY-1S effectively improved soil workability and significantly reduced soil mechanical strength, cutterhead torque, and excavation resistance, thereby enhancing shield tunneling efficiency and operational adaptability.

Blockage of drainage holes in deep-buried hydraulic tunnels can significantly alter the pore-water pressure distribution and the mechanical response of the lining, thereby threatening tunnel stability and operational safety. Compared with conventional methods that represent blockage by uniformly reducing the permeability of drainage holes, this study proposes a segmented blockage modeling method based on line elements. A numerical seepage model for deep-buried tunnels is then established and validated its good applicability through comparative analyses. On this basis, an indirectly coupled seepage–stress procedure is employed to develop a load–structure model and to quantify the influence of the spatial distribution of blocked drainage holes on the load-bearing behavior of the lining. The results show that hydraulic damage in the lining induced by single-hole blockage follows a three-stage evolution: gradual change, abrupt increase, and stabilization. The corresponding critical blockage depths are 0 m, 3.6 m, and 4.4 m. Excessive hydraulic gradients near the blockage cause compressive deformation of the lining lead to the formation of an inflection point. At the end of the gradual stage, when the relative blockage depth reaches 0.9, tensile damage governs the failure of the lining, whereas compressive damage can be neglected. For adjacent multi-hole blockage, local external water pressure on the lining increases significantly, reaching a peak of approximately 3.81 MPa when two blocked holes are adjacent and decreasing to 2.65 MPa when they are widely spaced. The resulting highly non-uniform local high-pressure gradient is the dominant factor causing through-thickness damage. These findings provide a theoretical basis for optimal design, safe operation, and effective maintenance of drainage systems in deep-buried hydraulic tunnels.

As the most widely adopted type of dam, earth-rockfill dams are widely recognized for their considerable deformation adaptability and cost-effectiveness. Settlement occurring during and after the construction of earth-rockfill dams can lead to serious hazards. Therefore, accurate prediction and control of dam-body displacement are crucial to construction safety. This paper presents a finite element method (FEM) analysis of the settlement behavior of the Altash dam in Xingjiang. The stress-strain characteristics of the rockfill are examined using both the unified hardening model for clays and sands (CSUH) and the modified Cam-clay (MCC) model. The model parameters are calibrated using a differential evolution algorithm integrated with a modified local search method. The results show that the CSUH model predicts settlement behavior significantly more accurately than the MCC model. This enhanced performance is attributed to the introduction of the parameter p_s, which enables a more comprehensive representation of rockfill compressibility. To further improve the accuracy of settlement prediction, this study proposes an optimization-based calibration method that accounts for the actual settlement behavior of dams and uses only triaxial test data within the strain range observed during construction. The rationality of proposed method is validated through comparison between numerical predictions and monitoring data.

In rock engineering scenarios, such as mining extraction, underground energy storage, and deep tunnel excavation, rock fracture behavior is often abrupt and potentially destructive. Therefore, accurately perceiving the different stages of rock fracture is essential for geological hazard early warning and engineering risk mitigation. However, how to achieve multi-source feature fusion and reliable state discrimination of rock fracture behavior based on existing monitoring techniques, especially acoustic emission, still requires further investigation. This study proposes an intelligent identification framework for rock fracture stages driven by the fusion of knowledge and data. Taking granite as a representative material, stage classification rule knowledge was first established based on rock mechanics theory, and the resulting rules were then used to automatically annotate acoustic emission data. On this basis, three machine learning algorithms, namely extreme gradient boosting (XGBoost), random forest (RF), and support vector machine (SVM), were employed in a data-driven framework to identify different stages of granite fracture. A comparative analysis of the three models shows that the XGBoost model achieves the best overall performance, with a weighted F1-score (the weighted average of per-class F1-scores by sample count) of 0.860 3, a macro-average recall of 80.16%, and a macro-average AUC of 0.97. Macro-average AUC refers to the arithmetic mean of AUC values across all classes. The receiver operating characteristic (ROC) curve plots false positive rate versus true positive rate to illustrate model discriminative performance at different thresholds. AUC denotes the area under the ROC curve, ranging from 0 to 1, where higher values correspond to better binary classification performance. The key interpretable features for stage recognition include cumulative energy, cumulative ring-down count, ring-down count, and hit count. Among these, cumulative energy and cumulative ring-down count showed the largest SHAP (quantify the contribution of each feature to model predictions) contributions. This study establishes a new paradigm for the automatic identification of rock fracture states and disaster warning in rock engineering.

 In traditional soil mechanics, the cohesion of clayey soils is often described using the macroscopic cohesion parameter c within the Mohr-Coulomb failure criterion. In this study, siloxane-coated glass particles were used to prepare controllable cohesive granular materials (CCGM), enabling quantitative control of the microscopic cohesive forces between particles, that is, polymer-bridge bonding. This approach links particle cohesion to microscopic interparticle bonding forces and shifts the description of particle behavior from a macroscopic framework to a quantifiable physical characterization. To reveal the cross-phase evolution mechanism by which cohesive particles evolve from a quasi-static state to a flowing state, this study introduces a dynamic cohesive-force expression and develops a  rheological model for cohesive particles that relates the frictional coefficientto the inertial number I. Based on the proposed model and the material point method (MPM) framework, a continuum-mechanics model was established to describe the full transition of cohesive particles from rest to flow. The continuum model effectively reproduces the temporal evolution process of particle column collapse and captures experimentally observed phenomena, including delayed collapse initiation, reduced runout distance, and an increased angle of repose with increasing cohesion. Experiments and simulations also show that when the cohesive number Co≥8.1 or the aspect ratio a≥2.0, the granular flow forms a radial crack network due to non-uniform energy dissipation. The research provides a theoretical framework for predicting landslide and debris-flow motion.

Perception and prediction of soft-ground consolidation behavior constitute a prerequisite for stability analysis and settlement calculation of soft foundations; accurate evaluation of the degree of consolidation hinges on the reliable determination of consolidation parameters and drainage boundary conditions. To identify consolidation parameters and consolidation characteristics of soft soils, a physics-informed neural network (PINN) is developed for the soft-foundation consolidation equation. The model uses the automatic differentiation capability of neural networks to represent the governing differential equation. Continuous-drainage boundary conditions are introduced to account for time-dependent pore-water pressure at the boundaries. A composite loss function integrating the physical equation, continuous-drainage boundaries, and initial conditions is constructed, while pore-water pressure dissipation data further incorporated as a data-driven term. Consequently, a physics- and data-driven model is established for the integrated analysis and prediction of soft-ground consolidation behavior. Through model training and hyper-parameter optimization, in-situ consolidation coefficients of soft soils and parameters of continuous-drainage boundaries are inversely identified, enabling the digital-intelligent prediction of consolidation behavior. After verifying the inversion and prediction accuracy of the model, the influences of piezometer layout, monitoring period, and acquisition frequency on parameter-identification accuracy are investigated. Results indicate that placing piezometers in locations with pronounced pore-water pressure variations markedly improves identification accuracy. Parameter identification accuracy also increases with longer monitoring periods and higher data acquisition frequencies. By effectively integrating physical mechanisms, the proposed model accurately identifies multiple unknown consolidation parameters using only short-term pore-pressure observations. It also provides integrated predictions of pore-pressure dissipation and consolidation development. This study offers a reliable and efficient approach for soft-ground treatment design, construction control, and post-construction settlement prediction.

Shrinkage and temperature-gradient action in self-compacting concrete can cause damage and deformation in ballastless track structures, resulting in irreversible degradation of the interfacial bonding performance between the track slab and the self-compacting concrete layer. In view of this, this study takes CRTS III ballastless track as the research object. Using the finite element method, a refined model that considers structural reinforcement and interlayer bonding is developed. The deformation behavior of the track structure and the damage evolution mechanism of the interlayer interface under thermal loading are systematically investigated. Combined with the multi-body system dynamics method, the vehicle-track coupling dynamics model is then established, and the influence of periodic track irregularity induced by temperature gradient action on the dynamic response of vehicle-track system is analyzed. The results show that shrinkage of the self-compacting concrete is effectively restrained by the door-shaped steel bars, which suppress the progression of internal interface damage. The interface damage between the track slab and the self-compacting concrete layer is mainly governed by positive temperature-gradient loading. When the gradient is +60 °C/m, the damage begins at the slab corner and gradually expands to the slab. Under a temperature gradient of +70 °C/m, localized failure occurs. In addition, temperature-gradient action induces periodic irregularities in the ballastless track unit, thereby intensifying wheel–rail dynamic interaction. The wheel–rail vertical force and track-slab vibration acceleration increase with increasing temperature gradient and vehicle speed.

Accurate determination of rock mechanical parameters is crucial for engineering design. While-drilling technology has become one of the primary approaches for assessing rock strength. Existing while-drilling strength prediction models typically rely on a single approach, such as theoretical physical analysis or machine learning. As a result, their predictive performance is highly sensitive to field operating conditions and data quality. This study develops a physics-informed neural network (PINN) model to predict the uniaxial compressive strength of rock from while-drilling parameters. The model is built on a CNN-LSTM-Attention fusion architecture. It achieves accurate prediction of rock uniaxial compressive strength by integrating physical knowledge with data-driven learning. Specifically, the physical model for strength prediction based on multi-wing cutting drilling is transformed into constraint conditions within the network. Based on parameter sample construction and comparative analysis, the modified physical model improves rock strength prediction accuracy by 15%. Its correlation coefficient with the measured rock strength reaches 0.97. Compared with the 6 traditional neural network models, the determination coefficient of the strength prediction results of the PINN model on the test set is increased by at least 2%, and the mean absolute error is reduced by at least 9%. Further validation using actual while-drilling data from composite formations shows that, when the lithological sequence is granite, conglomerate, and sandstone, the average prediction errors of the PINN model are 3.63, 4.44 and 1.19 MPa, respectively. In all cases, the percentage error is less than 8%, demonstrating the model’s strong robustness in predicting the strength of formations with different lithologies. These findings provide practical guidance for tunnel while-drilling technology and the real-time automated characterization of rock mass strength.

To accurately evaluate the time-varying reliability of reservoir slope, it is necessary to consider both the spatial variability of soil strength and hydraulic parameters and the temporal effects of environmental loads, such as rainfall and reservoir water-level fluctuations. In time-varying reliability analysis, traditional numerical methods often require static reliability to be evaluated at each time step, making the process computationally expensive. Therefore, a hybrid deep-learning-based method is proposed for the efficient analysis of the time-varying reliability of reservoir slopes. The proposed method incorporates a convolutional block attention mechanism (CBAM)-enhanced residual network (ResNet) to extract spatial features from random-field images of soil parameters. It also integrates a bidirectional long short-term memory network (BiLSTM) to capture the temporal evolution of environmental loads. The outputs of the two models are fused to achieve time-series prediction of the safety factor of reservoir slopes under the uncertainties of soil parameters and environmental loads, thereby realizing efficient reliability evaluation. The proposed method is validated using the Shiliushubao landslide in the Three Gorges Reservoir area as a case study. The results show that the proposed method can simultaneously learn from multiple random-field images of soil parameters and multimodal temporal data describing environmental loads. The predicted safety factors are consistent with those obtained using traditional numerical methods, while computational efficiency is improved by approximately 305 times. Using monitoring data spanning 0 to 60 months, the proposed method achieved a coefficient of determination (R2) of .932 for failure probability prediction, outperforming conventional deep learning approaches and demonstrating superior predictive capability.

Downhole acquisition of drilling parameters for geotechnical property analysis is a key component of intelligent engineering site investigation. To address the limitations of existing measurement-while-drilling (MWD) systems, which lack generalizability and cannot directly measure torque and thrust at the drill bit. A storage-based downhole MWD method is proposed. By introducing a measurement subassembly between the drill rod and drill bit, this method enables direct near-bit measurement of torque, thrust, rotational speed, and inclination. These measurements have clear physical significance and eliminate the need for rod-length corrections. The subassembly adopts a modular design, allowing flexible adaptation to different drill rod sizes and compatibility with various drilling rigs. This design ensures strong engineering applicability. The system independently collects data in a storage mode, while a depth counter performs data splicing and time–depth conversion. This process ensures data completeness and reliability. Field tests conducted at an industrial park in Wuhan demonstrate that the proposed method can clearly capture the drilling characteristics of soil and rock layers, as well as lithological transition zones and locations of discontinuities. Combined analysis of multiple parameters effectively reveals variations in rock integrity. Within the same lithology, the linear increase in thrust with drilling depth confirms that near-bit direct measurement mitigates rod-length effects. This technique provides a feasible, flexible, and high-precision approach for MWD data acquisition and geotechnical parameter inversion.