Deeper exploitation of coal resources faces higher possibility of rockburst and mining earthquake. Hydraulic fracturing (HF) proved to be an effective solution in coal mines but its monitoring and evaluation remains unexplored. This study presents a true triaxial hydraulic fracturing experiment with acoustic emission (AE) recording, conducted on a 150 mm cubic coal sample from a deep underground mine. Fracture evolution was analysed based on 539 localized events using a modified simplex method. According to the injection history, AE counts and energy evolution, the coal undergoes initiation, intersection, and breakdown stages. The AE energy peak lags behind the water pressure peak due to the dilatancy effect. Rapid fluctuations in water pressure trigger peak AE event rates, often followed by silent periods with minimal AE activity, which can serve as precursors for crack prediction. The distribution of AE events indicates that fractures originate from the naked borehole and propagate upward, predominantly accumulating in the middle and upper regions of the coal sample. At low water pressure, fractures extend primarily along the maximum principal stress direction, while at high water pressure, they diffuse spherically. The uneven transition and storage of crack energy during injection lead to alternating shrinkage and expansion of AE event distribution as water pressure increases. In addition, coal heterogeneity plays a significant role in fracture formation, resulting in tortuous hydraulic fracture planes that deviate from alignment with the maximum principal stress.

With mines reaching into greater depth, problems with violent failures become more prevalent, such as rockbursts. To assess the potential for such occurrences, understanding energy dynamics is crucial. This study delves into the impact of excavation shape, aspect ratios, inclined angles and in-situ stress on dissipated plastic energy and released kinetic energy through numerical simulations. Detailed analyses of plastic zone around rectangular and elliptical excavations, varying in aspect ratios, unveiled the distinct patterns of tensile and shear plastic zone. Quantitative assessments of dissipated plastic energy and released kinetic energy highlighted contrasting behaviors between rectangular and elliptical excavations across various aspect ratios. Although the relationships between these energies and excavation parameters are found to be intricate, it becomes evident that: with an increase in the aspect ratio, both shear and tensile plastic zones exhibit a growing trend, and their distribution shows a high degree of consistency with regions of energy concentration. When the aspect ratio is below 1.5, elliptical excavations release slightly less kinetic energy than rectangular ones. Beyond 1.5, this relationship reverses. Moreover, upon exploring inclined angles, critical angles were identified, delineating points where the influence of aspect ratio nearly diminishes. Under the given in-situ stress conditions, approximately 50° and 20° serve as the critical angles for dissipated plastic energy and released kinetic energy, respectively. Changes in the degree of in-situ stress anisotropy have a limited impact on the overall energy trends but significantly alter the critical values of aspect ratio and inclination angle. When the in-situ stress ratio matches the excavation aspect ratio, both forms of energy reach their minimum values. These findings illuminate the complex interplay between excavation geometry and energy dissipation, offering invaluable insights for designing effective excavations and devising strategies to mitigate failures.

To explore the evolution process and failure characteristics of rockburst in rock masses with a double free surface structure, we conducted a comparative study between true triaxial double-face rapid unloading rockburst tests and single-face rapid unloading rockburst tests, using a true triaxial experimental system capable of rapid unloading from multiple faces. The main conclusions are as follows: the maximum ejection velocity of double-free surface rockburst fragments is higher than that of single-free surface rockburst (20 %-40 %), and the rockburst ejection duration is generally longer than that of single-free surface rockburst; the peak stress of double-free surface rockburst is smaller than that of single-free surface rockburst, but there will be a more obvious unloading platform stage during unloading. Compared with single-surface unloading rockburst, double-surface unloading is more likely to cause significant brittle failure and violent rockburst; the AE events of single-free surface rockburst are concentrated near one free surface, while the AE events of double-free surface experiments are distributed near two free surfaces. In the later stage of both rockburst experiments, the number of shear cracks will exceed that of tensile cracks. The extra free surface of double-free surface rockburst will enhance crack extension, especially shear cracks, resulting in more serious rock damage.

The northern highlands of Ethiopia often experience slope failures triggered by rainfall. Landslides are common in Lesalso, Laelay Maichew Wereda, northern Ethiopia, causing destruction to homes, crops, agricultural lands, the death of wildlife, and the eviction of local residents from their homes. The primary objective was to identify the causative and triggering factors for the landslide occurrence. Field investigations, soil laboratory test, vertical electrical sounding (VES) and profiling were conducted. The landslide affected area is characterized by aphanitic basalt with small agglomeratic basalt, phyllite, laterite sandstone, metachert and rhyolite rock units. Physical and engineering properties of 22 soil samples indicated that the failure materials mainly consist of fine grain soils (72.72 %), liquid limit (29.5 %-66.4 %) and plasticity index (7.6 %-43.65 %), medium to high degree of swelling (0.08 %-71.5 %), variable low to high water content (21 %-54.3 %), specific gravity (2.45-2.81), dry unit weight (1.467 g/cm³ to 2.254 g/cm³), activity of soils (0.379 %-1.78 %), soil compression index (0.1785 %-0.43685 %), shrinkage index (0.27 %-28.61 %), shrinkage limit (0.062 %-30.86 %), friction angle (φ) (7.68°-47.94°) and cohesion (C) (6.45 kPa-52.72 kPa). The primary factors influencing slope stability include the severely weathered and impermeable rocks beneath the hard and fractured surface, the presence of fine grain soils (clay and silt), a geohydrological setting that facilitates the accumulation of water pressure within the slope, and a steep slope that can generate sufficient stress to induce failure. These findings provide critical insights for the development of proactive mitigation strategies to protect local communities and infrastructure from landslide hazards.

The identification of rock stability and the prediction of failure time are crucial for the early warning and prevention of sudden geological disasters such as landslides and collapses. To address these challenges, this study proposes three convolutional prediction models: CNN-LSTM-Attention, CNN-BiLSTM-Attention, and CNN-GRU-Attention. The displacement coordination coefficient (DCC) index and stress curves were employed as input variables to evaluate the performance of each model in discriminating rock stability states under different data structures and input configurations. Furthermore, an innovative methodology for predicting rock failure time utilizing convolutional models was developed. The experimental results demonstrate that the CNN-LSTM-Attention model, utilizing a 10 × 10 × 2 data structure, exhibits superior performance in rock stability state discrimination, achieving an accuracy of 95.25 % on the validation set and a recall rate of 96 % for samples in high-risk areas. Furthermore, when the DCC index was used as the input variable, the CNN-LSTM-Attention model achieved recall rates of 95.8 % and 86.5 % for medium- and high-risk areas, respectively, in the validation set. These findings indicate that the proposed convolutional models can effectively identify rock stability states by leveraging surface deformation characteristics. The CNN-LSTM-Attention model, with the DCC index as the input variable, is capable of predicting the final rock failure time in real-time once the DCC abrupt change exceeds 0.78. For different rocks, the model can predict the failure time within 20 s after the DCC reaches 0.78, with an error rate of less than 9 %. The convolutional neural network model, developed based on the DCC index, provides a novel methodological approach for geohazard early warning research, facilitating slope instability monitoring and earthquake precursor identification using GNSS and other displacement measurement techniques.

Recurring and frequent landslides in the Western Ghats of India pose major socio-economic challenges by interrupting the infrastructure development and daily life. The present study evaluates the landslide susceptibility of the region using a hybrid approach. Data from various sources including satellite images, past landslide records, geological, topographical and hydrological datasets was utilized to develop landslide inventory and the causative factors of the study area. The adopted hybrid approach integrates the analytic hierarchy process (AHP) and relative frequency ratio. The model demonstrated excellent discriminatory ability with an area under the ROC curve (AUC) of 0.902. The Uttara Kannada district has been identified as most susceptible to landslide in Karnataka. Further, the recent Ankola landslide in the highly susceptible Uttara Kannada was taken for a detailed case study. The field observations, geotechnical characterization, and rainfall details of the landslide site suggest that the anthropogenic activity and heavy rainfall were the reasons for triggering the landslide. The landslide event was back-analysed using pore water pressure factor (R_u) to simulate pore-water pressure development during rainfall. The R_u factor of 0.4 was identified as critical threshold for initiating slope failures, providing a quantitative basis for early warning systems for the region. The susceptibility model and the R_u factor threshold found in the study offer essential information for the enhancement of risk mitigation efforts in the Western Ghats.