Accurate fault diagnosis of mining hoisting head sheave systems is critical for ensuring operational safety in harsh underground environments. This study proposes a multi-source fault diagnosis framework that fuses vibration and acoustic information using a Convolutional Neural Network and Random Forest (CNN-RF). To support mechanism understanding and validate the experimental platform, finite element and multi-body dynamics simulations (ANSYS/ADAMS) are employed for physical verification and fault signature analysis, while the CNN-RF model is trained and tested exclusively using experimentally acquired vibration and acoustic data. For feature construction, vibration signals are transformed into time–frequency representations (including STFT, CWT, and generalized S-Transform (GST)), and acoustic signals are characterized using Mel-Frequency Cepstral Coefficients (MFCCs). Experimental results demonstrate that vibration–acoustic fusion improves diagnostic performance compared with single-modality baselines; the best performance is achieved by GST+MFCC with the proposed CNN-RF classifier, reaching an accuracy of 98.96%. Future work will conduct cross-condition validation under varying speeds and loads and investigate missing-modality robustness to further assess generalization and deployment reliability.

Accurate characterization of underground geological conditions is essential for gas control, geological hazard assessment, and safe coal mining operations. However, conventional geological interpretation methods often suffer from limited spatial accuracy due to borehole deviation, sparse geological control, and insufficient integration of multi-source borehole data. To address these limitations, this study proposes an integrated geological characterization framework combining resistivity-based image logging, borehole trajectory correction, and BIM-based three-dimensional geological modeling using 135 gas extraction boreholes from the Coal Seam 15-21050 working face of Pingdingshan No. 8 Coal Mine, China. Multi-parameter logging data, including natural gamma, apparent resistivity, natural potential, and borehole image observations, were used to identify coal seam lithology, stratigraphic interfaces, and structural characteristics. Borehole trajectory analysis revealed systematic deviation patterns controlled by borehole inclination, lithological heterogeneity, and drilling conditions, highlighting the necessity of trajectory correction for accurate spatial positioning. Trajectory-corrected borehole coordinates were subsequently integrated into a BIM-based three-dimensional geological reconstruction workflow using spatial interpolation methods. The resulting model successfully reproduced coal seam geometry, interburden distribution, and localized concealed structural anomalies. Coal Seam 15 exhibited thicknesses ranging from 2.69 to 3.47 m, while Coal Seam 16–17 ranged from 1.51 to 2.38 m. The proposed workflow improved the reliability of geological interpretation and the accuracy of spatial characterization, providing an effective technical basis for gas drainage optimization, geological hazard assessment, and intelligent underground coal mining.

Hydroclimatic stress in mining-affected landscapes is influenced not only by rainfall but also by abrupt rainfall transitions and their interaction with subsurface water conditions. This study examines rainfall-transition extremes and groundwater response in the Barapukuria–Parbatipur coal-mining region of northwestern Bangladesh using monthly rainfall and monthly simulated groundwater depth for 1981–2025. Rainfall transitions were standardized within calendar-month classes to derive a Standardized Transition Anomaly Index (STAI), allowing anomalous wetting and drying to be distinguished from normal monsoonal progression. The analysis identified 146 high-transition months at [[EQUATION]], including 72 positive and 74 negative anomalies, indicating that abrupt rainfall reversals are recurrent features of the regional hydroclimatic regime. Simulated groundwater depth showed a distinct seasonal structure, with the shallowest climatological condition in May at 3.69 m and the deepest in November at 7.63 m, demonstrating that groundwater seasonality is related to, but not synchronous with, rainfall seasonality. At the monthly scale, groundwater depth was significantly associated with rainfall and more strongly with raw rainfall transition, but not with same-month standardized anomaly. At the annual scale, rainfall showed only a weak, non-significant decline, whereas annual mean simulated groundwater depth deepened significantly at about 0.03 m year−1, indicating that long-term groundwater change cannot be explained by rainfall variability alone.