Rock burst is a serious threat to mine safety production, and its prediction is of great significance to effectively prevent and control the dynamic disaster of rock burst. Therefore, this paper used AE data of rock loading process to conduct pseudo-prospective prediction and short-term and long-term prospective prediction of rock failure, and further explore the prediction of rock burst based on acoustic emission (AE) and its reliability. The results show that: by selecting the appropriate critical point of failure, the autoregressive integrated moving average model can make short-term predictions of rock failure. The prediction accuracy of the acoustic emission posi- tioning technology for the fracture surface and fracture location of rocks is affected by the prediction time. The closer to the failure point, the higher the prediction accuracy is. The energy prediction method based on the energy accumulation mechanism can effectively predict the elastic energy at the moment of failure. This study also proposes combining machine learning with the analysis of historical acoustic emission data from rockbursts, which can improve the reliability of rockburst prediction. The research results can provide theoretical support for the prevention and control of rock burst dynamic disasters.

Brittle rocks exhibit significantly lower dynamic direct tensile strength compared to their compressive strength, and the tensile strength is relatively difficult to be quantitatively measured through experiments. While extensive research has characterized dynamic tensile behavior through indirect testing methods, the direct tensile strength remains critical for evaluating rock fracture mechanisms and ensuring the safety of deep underground engineering systems. Notably, the microcrack propagation dynamics governing dynamic direct tensile fracture in brittle rocks remain understudied. To address this gap, we develop a micro-macro dynamic tensile fracture model that elucidates the stress-strain constitutive behavior of brittle rocks under dynamic loading. The model integrates four key components containing the quasi-static microcrack growth kinetics, the microcrack length-macroscopic strain relationships, the crack growth rate-strain rate coupling, and the transition from quasi-static to dynamic fracture toughness. A critical strain rate εʹ1c causing the crack initiation stress to be the peak strength is investigated. Parametric investigations quantify the influence of crack extension rate on stress-crack length relation, strain rate on stress-strain relation, and the governing parameters (initial damage D0, microcrack size a, inclination angle φ, and density Nv) on dynamic crack initiation thresholds, peak strength and critical strain rate. Its validity is rigorously verified through comparative analysis with experimental data. The results will have significance for disaster evaluation in rock engineering.

The uniaxial compressive strength (UCS) of rocks is a crucial indicator for evaluating the bearing capacity of geological structures in rock engineering, and it holds significant implications for disaster management. How- ever, direct measurement poses a significant challenge. Therefore, simpler alternatives such as Schmidt hammer rebound number (SRn), P-wave velocity (Vp), and point load index (Is) are frequently used to estimate UCS indirectly. In this study, we compiled a comprehensive dataset of 1168 samples that included SRn, Vp, Is, and UCS values. The dataset was refined using an isolation forest algorithm, which identified and removed 280 outliers, leaving a dataset of 888 samples for analysis. We developed and assessed an automated machine learning (AutoML) model for predicting UCS, introducing a novel approach to tackle this prediction challenge. Additionally, we compared models enhanced by Bayesian optimization, including multi-layer perceptron (MLP), support vector machine (SVM), Gaussian process regression (GPR), and K-nearest neighbor (KNN). Among these, the AutoML model demonstrated superior performance in UCS prediction, offering a rapid and efficient method for estimating UCS in engineering applications and enabling intelligent classification of rock masses. The study also evaluated the sensitivity and contribution of SRn, Vp, and Is in UCS estimation by various techniques, including permutation feature importance (PFI), SHapley Additive exPlanations (SHAP), and local interpretable model-agnostic explanations (LIME). The results underscore that the AutoML approach not only streamlines UCS modeling but also provides a robust and comprehensive solution, significantly enhancing the accuracy and ef- ficiency of the prediction process.

Groundwater flow in fractured rock masses, governed by discrete fracture networks (DFNs), critically impacts tunnel engineering safety. This study addresses water-inrush disasters by proposing a DFN-based grouting optimization method for jointed rock masses (Grades II-IV). The structural grid model is used to evaluate the degree of rock penetration in this area. Permeability coefficients and the radii of permeability ellipses are calculated at 30-degree intervals along the network, enabling comprehensive evaluation. Utilizing the least squares method, seepage ellipses are fitted to determine primary seepage coefficients. In consideration of the most unfavorable scenarios, rock mass seepage coefficients are selected for grouting design calculation. For each grade of surrounding rock mass, assessments are conducted to ascertain the water inflow of unlined tunnels, the water inflow of lined tunnels, and external water pressure on tunnel linings. Tunnel curtain grouting is required when the tunnel water inflow exceeds the design limits. Appropriate parameters for grouting ring thickness and permeability coefficients are selected to fulfill engineering specifications. In cases of excessive external water pressure in tunnel linings and significant inflow of water into the tunnel, it is recommended that grouting and lining operations are carried out after drainage and pressure relief in the tunnel. The DFN methodology enables targeted grouting that reduces water-inrush risks in high-risk zones.

In the 21st century, climate change has exacerbated global instability, leading to a rise in landslide occurrences. In Bangladesh, mountainous areas such as Bandarban experience significant landslides during the monsoon season. This study seeks to evaluate landslide susceptibility in Bandarban and identify hotspots for optimal landslide hazard mitigation. This study examined landslide susceptibility using the analytical hierarchy process (AHP) and spatial weighted overlay (SWO). Ten conditioning factors were considered, with AHP based on re- sponses from 100 key respondents. Using field surveys and high-resolution satellite images, 280 landslide occurrence samples were collected to rank the subfactors. Using AHP-derived weights of factors and subfactors, the SWO approach was used to create the landslide susceptibility map (LSM). The Getis-Ord (Gi*) spatial sta- tistics was then used to generate landslide susceptibility hotspots. The result showed that human influence weight 17.02%, making it the most crucial factor in landslide susceptibility. AHP-derived weights were reliable because their consistency ratio was <0.1. According to the study, 59.86% of the area is moderately susceptible,20.06% is high, and 4.31% is very high. The validation of LSM by ROC curve found excellent performance (AUC = 0.93) of the approaches. Specifically, 63.8% of very high susceptibility areas and 33.26% of high susceptibility areas were found within the hotspot zones with 99% confidence. The research showed the combined use of field samples and remote sensing-based spatial variables improved the accuracy of LSM. These findings can be useful for ensuring proper land use planning and implementation of landslide hazard mitigation measures.

In underground coal mining, surface subsidence disasters are likely to be induced. Especially, under the condition of multi-seam mining, the movement characteristics of the overlying strata are more complex. Once these characteristics are transmitted to the surface, it is easy to lead to intensified deformation and the appearance of ground fissures. This not only causes damage to surface buildings but also may have irreversible impacts on the aquifer. Taking 1208# working face of Hongyang No. 3 Coal Mine as a case in study, this paper uses the Small Baseline Subset Interferometric Synthetic Aperture Radar (SBAS-InSAR) technology to systematically monitor and analyze the surface subsidence characteristics of the multi-mining area (MMA) and the single-mining area (SMA) changing over time, with a focus on discussing the subsidence laws of the MMA. The comparative analysis results show that: (1) There is an obvious hysteresis in surface subsidence, the position of the subsidence center basically corresponds to that of the working face, but the influence range of subsidence exceeds the boundary of the working face, besides, significant surface subsidence occurred 36 days after mining the No. 1208 working face, and the change in the structure of the overlying strata was transmitted to the surface; (2) Compared with the single-mining area (SMA), the maximum subsidence rate (MSR) and the maximum subsidence value (MSV) in the multi-mining area (MMA) are higher, and both the subsidence center and its influence range are significantly expanded; and (3) After the mining of the working face stops, the subsidence rate slows down, but the subsidence increment in the MMA area is still higher than that in the SMA. The above findings deepen the understanding of the evolution mechanism of surface subsidence disasters caused by multiple mining activities, and provide an important basis for the monitoring, prevention and control of subsidence disasters in similar mining areas.