Rock excavation is essentially an unloading behavior, and its mechanical properties are significantly different from those under loading conditions. In response to the current deficiencies in the peak strength prediction of rocks under unloading conditions, this study proposes a hybrid learning model for the intelligent prediction of the unloading strength of rocks using simple parameters in rock unloading tests. The XGBoost technique was used to construct a model, and the PSO-XGBoost hybrid model was developed by employing particle swarm optimization (PSO) to refine the XGBoost parameters for better prediction. In order to verify the validity and accuracy of the proposed hybrid model, 134 rock sample sets containing various common rock types in rock excavation were collected from international and Chinese publications for the purpose of modeling, and the rock unloading strength prediction results were compared with those obtained by the Random Forest (RF) model, the Support Vector Machine (SVM) model, the XGBoost (XGBoost) model, and the Grid Search Method-based XGBoost (GS-XGBoost) model. Meanwhile, five statistical indicators, including the coefficient of determination (R²), mean absolute error (MAE), mean absolute percentage error (MAPE), mean square error (MSE), and root mean square error (RMSE), were calculated to check the acceptability of these models from a quantitative perspective. A review of the comparison results revealed that the proposed PSO-XGBoost hybrid model provides a better performance than the others in predicting rock unloading strength. Finally, the importance of the effect of each input feature on the generalization performance of the hybrid model was assessed. The insights garnered from this research offer a substantial reference for tunnel excavation design and other representative projects.

In underground coal gasification (UCG) projects, the flow of high-temperature gas leads to the simultaneous exposure of surrounding rock in tunnels to high-temperature and cyclic load, posing risks of rock instability. To investigate the failure modes and precursors of rock instability under high-temperature and cyclic loading, uniaxial compression tests and cyclic loading–unloading tests were conducted, utilizing acoustic emission (AE) technology and discrete element method (DEM). In addition, the microstructural damage of red sandstone after heat treatment was analyzed utilizing scanning electron microscopy (SEM) images. The test results reveal that increasing temperature leads to a delayed appearance of change points in AE counts and b value and more pronounced AE signals. Fatigue loading cannot always cause severe damage to red sandstone at the first cycle of the load level, according to the 25 °C group, 400 °C group, and 600 °C group specimens. RA-AF data distribution reveals that shear crack proportion generated by uniaxial compression in red sandstone subjected to the same temperature treatment is significantly greater than that caused by fatigue loading. Furthermore, the shear crack proportion in red sandstone specimens drops and then climbs with increase in temperatures. Based on the critical slowing down (CSD) theory, this paper analyzes the variance curves and autocorrelation coefficient curves of multiple AE parameters. The sudden drop point of the b value was taken as the starting point of rapid crack propagation, and the mutation points of various AE parameters lagging behind this starting point were searched for. It was found that the change point of AE energy variance is closest to the specimen instability point, which can be employed as a precursor for specimen instability. Heat treatments cause the internal structure of the red sandstone to evolve from a pompon-like structure to a flocculated porous structure. In addition, thermally induced cracks gradually develop and connect.

Previous joint test studies have mainly been conducted under the condition of uniformly distributed loads, but in engineering, overlying loads are often non-uniformly distributed. It is necessary to investigate the shear mechanical properties of joints under non-uniform loads. This paper establishes three typical mechanical models: far-field concentration (FFC), near-field concentration (NFC), and focus on center (FOC). Direct shear test simulations using the DEM software PFC reveal that the location of load concentration affects main shear parameters. The closer the load concentration is to the far end along the shear direction, the greater the deformation of rock mass during failure is, the higher the proportion of shear cracks is, and the greater the strength and energy required for failure are. Compared to a uniformly distributed load (UD) condition, FOC shows similar mechanical properties; NFC yields inferior outcomes, while FFC results in superior mechanical performance compared to UD. By utilizing machine learning, five prediction models for non-uniform load shear strength are developed, with the Genetic-XGBoost algorithm demonstrating the highest accuracy. Weight calculation results indicate that load distribution form is the most critical factor influencing shear strength.

Jointed soft-hard composite rocks are frequently encountered in nature, and this complex structure contributes to unpredictable fracturing mechanisms and failure behavior. In this study, soft-hard composite rocks with three joints were fabricated to conduct a uniaxial loading experiment, supplemented by Digital Image Correlation (DIC) and Acoustic Emission (AE) experiments. The results indicate that the mechanical parameters display a V-shape variation trend with the increase of joint angle, which minimized at 30°. The peak strength ranges from 33.48 MPa to 44.93 MPa. The failure characteristics change from tensile failure to shear failure and finally to intact failure. According to the displacement curves on both sides of the crack, the initiation of wing cracks is driven by the direct tensile displacement field and indirect tensile displacement field for specimens with joint angles of 0–30° and 75–90°, respectively. While the crack initiation from joint tips corresponding to specimens with a joint angle of 45–60° is controlled by direct and indirect tensile displacement fields. Wherein the cracks initiate from the coplanar joint in the hard layer, driven by the indirect tensile displacement field, and the cracks expanding upward from other joint tips are more susceptible to the indirect tensile displacement field.

Jointed composite rock masses are ubiquitous, where the interplay between fractures and bedding planes significantly complicates their failure mechanisms. This study conducted uniaxial compression tests on a composite rock material featuring soft and hard components with unparallel double joints. Leveraging the digital image correlation (DIC) method and acoustic emission (AE) technology, the investigation analyzed the influence of joint angles on the mechanical properties, failure modes, and crack propagation behaviors. The results indicate that the peak strength exhibits an ascending–descending–ascending trend with the variation of upper joint angle, peaking at 90°. The failure modes can be systematically categorized into three types. The failure of the specimen is predominantly influenced by the upper joint. However, when the lower joint angle is relatively flat, it becomes more prone to contributing to the failure of the specimen. The lower joint-dominated failure mode only occurs when the upper joint angle is 90°. As the lower joint angle remains at 90°, an augmentation in the upper joint angle initially leads to a decrease and subsequently to an increase in the proportion of shear cracks. In contrast, when the lower joint angle deviates from 90°, the increase of the upper joint angle initially boosts and then diminishes the proportion of shear cracks.