Xiaofei Liu

Against the backdrop of frequent extreme short-duration heavy rainfall events, the precursors to sudden ground subsidence in the overlying strata of shallow-buried tunnels are often subtle, and the relationships between multiple field responses remain unclear; consequently, traditional single-field threshold-based early warning systems struggle to balance early detection with reliability. To address the challenges of multi-source, heterogeneous monitoring and graded early warning for ground subsidence in shallow-buried tunnels induced by heavy rainfall, a risk-state judgement framework integrating statistical learning and mechanistic constraints has been developed. Based on networked parallel electrical resistivity tomography (ERT), fibre Bragg gratings (FBGs) and digital image correlation (DIC) physical experiments, the temporal rates of change and spatial evolution characteristics of the three-field responses were organised into six-dimensional inputs. Following baseline normalisation, these were fed into a random forest algorithm, with ensemble voting used to characterise the relative support levels for four state categories; an asymmetric misclassification cost was introduced, and the initial level was determined by minimising conditional risk, with the final output constrained by mechanism-triggered risk thresholds and hysteresis rules. The results indicate that, in this experiment, representative precursors related to electrical properties, internal strain and macroscopic deformation appeared in sequence, preceding the collapse by 2157, 1203 and 166 s respectively. Based on the existing sample, the internal validation accuracy was 96.5 per cent, the macro-average F1 score was 94.7 per cent, and the recall rate for the “red” level was 91.3 per cent; the global feature contributions of FBG, ERT and DIC were 0.42, 0.37 and 0.21, respectively. Equivalent hydromechanical calculations provide a supplementary explanation for the relationship between seepage, stress adjustment and deformation development. This framework maps multi-field characteristics onto discrete early-warning levels that balance the cost of misclassification with physical conditions. The proposed method transforms the objective limitation of macroscopic precursor lag into the necessity for early, coordinated multi-field early warning, thereby providing a mechanistically sound and transferable graded early-warning framework for shallow-buried underground structures under heavy rainfall conditions.

To investigate the influence of joint spatial orientation on the rockburst mechanism of deep roadways, this study employs a true triaxial rockburst testing system to conduct physical simulation experiments on three types of limestone specimens from Yunnan: joint-free, parallel-jointed, and vertical-jointed. By integrating multi-source data acquisition techniques including high-speed imaging, acoustic emission (AE) monitoring, and fragments sieving, this research systematically analyzes the macroscopic failure process of rockburst, the evolutionary patterns of AE signals, and the characteristics of fragment generation under different fracture configurations. Joint orientation critically governs the rockburst failure mode, with parallel joints promoting extensive spalling and vertical joints localizing damage. Parallel joints act as guiding weak planes, resulting in a wider and more intense failure range, while vertical joints cause localized failure. The characteristics of acoustic emission signals reveal the regulatory mechanism of joints on the AE energy accumulation and release process. The parallel-jointed specimens exhibit a “step-like” increase, reflecting their phased damage and stress redistribution. Quantitative analysis of fragments indicates that the total mass of fragments in the parallel joint specimens is the largest, and the fractal dimension is the highest, indicating the most thorough fragmentation. This research provides crucial experimental evidence for the precise identification and differentiated prevention and control of rockburst risks in deep jointed rock mass engineering. It should be noted that these findings are derived from exploratory laboratory comparisons based on one specimen for each joint configuration, and further repeated tests are required to verify the generalizability of the observed patterns.

Deep underground engineering disasters stem from the cross-scale evolution and catastrophic failure of rock masses, with the occurrence and severity of such disasters being contingent upon the dynamics of crack propagation and energy release characteristics during the post-peak stage. To fundamentally elucidate the black-box processes involved in post-peak crack propagation and fracturing mechanisms in loaded rock, this study conducted mode-I fracture tests on quartzite beams under post-peak cyclic loading conditions. Through acoustic emission (AE) analysis, we investigated the spatiotemporal evolution of microcracks and developed a quantitative inversion method for microcracking source parameters based on forward modeling of the entire process, from source generation and wave propagation to signal reception. The evolution patterns of critical microcrack parameters, including type, scale, orientation, and damage, were systematically revealed, with mesoscopic inversion results validated against macroscopic measurements. Key findings include: During post-peak cycling, most microcracking events occur between the pre-peak damage stress and the post-peak stress drop period of the loading stage, with large-magnitude events concentrating in the upper-middle region of the crack. The AE Felicity ratio progressively decreases with the number of cycles, indicating weakening spatiotemporal memory effects. At the mesoscale, individual microcracks predominantly exhibit mixed-mode mechanisms. However, the global volume decomposition of microcracks shows that mode-I tensile components account for over 90% of the accumulated damage volume, where large-magnitude microcracks demonstrate horizontal tensile motion aligned with macroscopic fracture planes, confirming tensile-dominated mechanisms in three-point bending tests. Damage tensor analysis based on microcrack orientation and size reveals that preferentially oriented microcrack clusters dominate macroscopic fracture propagation. The increasing ratio between preferential and random damage components serves as a precursor for fracture progression. Additionally, the scale of microcracks and the energy dissipation range were quantitatively estimated. The average size of microcracks was found to be consistent with grain sizes, while the cumulative opening displacements of microcrack clusters exhibited a linear positive correlation with macroscopic crack opening displacement, thereby confirming the reliability of the inversion results. This research provides a quantitative methodology for representing rock fracture morphology, offering significant implications for understanding rock failure mechanisms and enhancing disaster prevention in rock engineering.

Implementing precise and advanced early warning systems for rock bursts is a crucial approach to maintaining safety during coal mining operations. At present, FEMR data play a key role in monitoring and providing early warnings for rock bursts. Nevertheless, conventional early warning systems are associated with certain limitations, such as a short early warning time and low accuracy of early warning. To enhance the timeliness of early warnings and bolster the safety of coal mines, a novel early warning model has been developed. In this paper, we present a framework for predicting the FEMR signal in deep future and recognizing the rock burst precursor. The framework involves two models, a guided diffusion model with a transformer for FEMR signal super prediction and an auxiliary model for recognizing the rock burst precursor. The framework was applied to the Buertai database, which was recognized as having a rock burst risk. The results demonstrate that the framework can predict 360 h (15 days) of FEMR signal using only 12 h of known signal. If the duration of known data is compressed by adjusting the CWT window length, it becomes possible to predict data over longer future time spans. Additionally, it achieved a maximum recognition accuracy of 98.07%, which realizes the super prediction of rock burst disaster. These characteristics make our framework an attractive approach for rock burst predicting and early warning.