The progressive failure of granular geotechnical materials acts as a critical precursor to landslide occurrence. To explore the feasibility of acoustic emission (AE) for early warning of such failures, direct shear tests were performed under eight shear rates to reproduce landslides with different deformation rates. AE signal analysis reveals that energy, ring down count, b-value and spectral centroid (Sc) evolve systematically throughout shearing. Specifically, the spectral centroid redshifts from 350 to 250 kHz, marking the development of macroscopic fractures. Integrating these AE signatures with shear stress evolution, the shear process is categorised into four stages: densification, elasticity, strain hardening and residual strength. Using 11 AE features, a convolutional neural network-long short-term memory hybrid deep-learning model embedded with spatial attention is constructed to automatically recognise shear stages, reaching a prediction accuracy of 94.52%. The attention module emphasises pre-failure key features and enhances model performance and interpretability. It is demonstrated that AE monitoring combined with deep learning can effectively identify precursor signals of shear failure in granular media under laboratory direct shear conditions, providing a preliminary methodological reference for the future development of landslide early warning systems.
The coupled effects of intermediate principal stress and matric suction on the mechanical properties of expansive soils remain poorly understood. To address this, an improved suction-controlled true triaxial apparatus was used to conduct suction equilibration tests, constant-suction isotropic consolidation tests and drained shear tests with various b-values on intact medium expansive soil, investigating the coupled effects of matric suction and intermediate principal stress on the mechanical behaviour of expansive soils. The results show that the intact expansive soil exhibits elastic–plastic volumetric deformation during wetting and drying. The yield suction sp = 71 kPa: elastic swelling occurs for 0 kPa ≤ s ≤ 37 kPa, elastic shrinkage for 37 kPa < s ≤ 71 kPa and plastic shrinkage for s > 71 kPa. The isotropic compression yield stress increases with suction, confirming the existence of the loading-collapse yield line. Dilatancy increases with suction and decreases with increasing intermediate principal stress. Cohesion increases non-linearly with suction and b-value, while the internal friction angle decreases with increasing intermediate principal stress. These findings provide a reasonable basis for stability assessments of expansive soil engineering under three-dimensional stress conditions.
Microscopic fracture evolution and macroscopic energy release were linked in weakly metamorphosed slate and metasandstone under triaxial confinement. Specimens were tested at 5 MPa, 15 MPa, 20 MPa and 25 MPa with synchronous acoustic-emission (AE) monitoring. Scanning electron microscopy images obtained before and after failure were processed using a validated fracture-extraction workflow, and the fractal-dimension difference (ΔDf) quantified changes in fracture-network complexity. AE-energy and pressure variables were integrated into a hybrid Bayesian non-linear regression–Gaussian process regression model. Retrospective threshold analysis showed that the anomalous cumulative AE energy at 20 MPa reflected numerous low-amplitude hits, few high-energy events and loading/contact disturbances; these data were excluded from model fitting. Both lithologies had positive ΔDf at 5 MPa and 15 MPa, whereas negative ΔDf at 25 MPa coincided with fracture closure, delayed failure and lower AE peak energy. The energy-peak and pressure ratios captured the non-linear dependence of ΔDf on energy release and confinement. The hybrid model outperformed the individual models (R2 = 0.826). ΔDf links microscopic fracture evolution to macroscopic AE response and supports pressure-dependent damage prediction within the tested range.
The mechanical strength of mudstone after absorbing water significantly decreases, seriously affecting the stability of tunnel projects. It is very important to study the laws and mechanism of this strength deterioration of mudstone. In this study, the main causes of the diseases in mudstone tunnels were determined. Through a series of the uniaxial compression and scanning electron microscopy tests, the strength deterioration mechanisms were studied and a statistical damage constitutive model was established. The results show that softening and expansion of mudstone after absorbing water were the main causes for the mudstone tunnel diseases. The probability of disease occurrence in mudstone tunnels from high to low was the inverted arch, vault, side wall and haunch. The softening and expansion of the mudstone were mutually reinforcing, and their mutual interaction formed a vicious cycle of water erosion. The stress–strain curve of mudstone showed a plastic–elastic–plastic curve. As the water content increased, the failure mode shifted from splitting failure to shear failure, while the degree of plastic deformation increased. Based on the Weibull distribution model and the Mohr–Coulomb failure criterion, a statistical damage constitutive model for mudstone absorbing water was established; the model can depict the failure process of softening mudstone.
