Fault prediction is critical for safe coal mine production. Traditional machine learning methods suffer from poor prediction accuracy when fault features are subtle. This study therefore proposes a WT-U-Net model by combining wavelet transform (WT) with U-Net to improve interpretation accuracy. Fault-related attributes were extracted from post-stack seismic data, where four attributes with low mutual dependence were identified through correlation analysis. The decomposition-reconstruction errors and energy differences of different wavelet basis functions applied to seismic data were then compared. The coif3 mother wavelet was selected for fault detection as its wavelet transform amplified fault-related signatures. The U-Net model was constructed to predict faults in the study area. Results demonstrate that the WT-U-Net model showed higher prediction accuracy than UNet alone on real datasets, with outputs more consistent with manual interpretations and improved convergence. The model also exhibited robustness and generalization in blind tests across other regions. The application of wavelet transform in seismic data denoising enhances fault-related signals, thereby increasing the model's accuracy. This study offers a new solution for intelligent fault identification in coal mining applications.

Wave velocity dispersion and energy attenuation are key parameters for precise coal exploration and development. Low-frequency stress-strain testing provides a strategy for realizing cross-band measurements, such as seismic bands. This study implemented a series of optimization to address the challenges of high signal-noise interference and high testing accuracy requirement in such test systems. Using four coal samples from Shenfu and Linxing areas, we measured and analyzed the dispersion and attenuation characteristics of coal rock seismic waves in the low-frequency band (4-1000 Hz) under varying temperatures and confining pressures. Experimental data were interpreted using the Chapman model to discuss the factors influencing these characteristics. Results show that: ① System performance was significantly enhanced by employing the Finite Impulse Response(FIR) bandpass filter and Fast Fourier Transform(FFT), which improved signal-to-noise ratio and the accuracy of phase difference extraction. ② The elastic parameters of coal rock increased nonlinearly with confining pressure. Both P- and S-wave velocities rose with increasing confining pressure, but decreased slightly with rising temperature. The attenuation peak decreased with higher confining pressure, and increased with temperature, while the eigenfrequency remained unchanged. The Chapman model shows a good fit with the measured data, confirming its applicability. ③ The attenuation peak was positively correlated with the porosity and water saturation of coal rock, with porosity exerting a more significant influence. Variations in fluid viscosity, permeability and water saturation were related to relaxation time, which is the main cause of dispersion and attenuation eigenfrequency.

The elasticity and physical properties of coal and rock, along with their variation characteristics, underpin the prediction of coalbed methane reservoirs and delineating "sweet spots". Physical properties are significantly influenced by temperature and pressure. This study aims to address the unclear dynamic response mechanism of pore-fracture structures in deep coalbed methane reservoirs under coupled in-situ temperature and pressure conditions, coal and rock samples were taken from the Linxing Block in the Ordos Basin. Variable temperature and pressure ultrasonic pulse transmission tests (0~30 MPa, 11~99 ℃) were conducted to analyze the response patterns of wave velocity and waveform characteristics to pore-fracture evolution. An improved dual-pore model was employed to quantify the temperature-pressure coupling constraint mechanism, establishing a nonlinear mapping relationship between waveform similarity coefficient and porosity. Results indicate: Waveforms exhibited significantly higher sensitivity to confining pressure than wave velocity. When confining pressure increased from 0 to 30 MPa, both P- and S-wave shape similarity coefficients changed by over 60%, representing a 2.96-fold~5.99-fold increase relative to wave velocity. As confining pressure increased, coal-rock pore closure primarily resulted from the synergy effects of rigid pore linear compression and flexible pore exponential closure. Temperature elevation (62℃, 19 MPa) increased the total porosity of coal samples by 0.55% compared to 11℃, 19 MPa conditions through thermal expansion effects. We obtained experimentally the waveform similarity coefficient-porosity mapping model (R²=0.786) and porosity inversion constraints (waveform, dual-pore, temperature-pressure), subsequently establishing a dynamic temperature-pressure, dual-pore fracture model for multi-parameter prediction of coalbed methane "sweet spots". These findings provide theoretical references for refined characterization and efficient development of coalbed methane reservoir properties.

Identifying the type, location and distribution of geological bodies in advance is crucial to ensure safe and efficient coal mining. Referring to the limitations of traditional 2D ground-coupled Ground-Penetrating Radar (GPR) in determining the direction and orientation of geological anomalies such as fault fracture zones, along with the significant safety risks associated with close-range detection near the working face in underground coal mines. A spatial scanning 3D GPR method is proposed for determining the position and orientation of concealed hazardous geological structures ahead. Firstly, by drawing on the theoretical model of fault fracture zone approximating to actual conditions, the forward numerical simulation was carried out using the finite difference method to analyze the time-domain response of fault fracture zone under different filling conditions. Secondly, a case study was performed on the Lvtang coal mine with complex geological conditions. Air-coupled 3D GPR was used to repeatedly collect multi-angle and multi-directional data from the +1730 horizontal transportation roadway of the mine. The advanced detection results were interpreted through time-profile analysis and comparative analysis of vertical and horizontal slices. Finally, the predictions for the orientation of hidden disaster-causing geological body were validated based on drilling and field exposure data. Results show that the underground application of air-coupled 3D GPR could accurately identify the geological body of the fault fracture zone within a 30-meter range in front of the mine. Filtering and wavelet transform techniques could extract valid wave information from signals with low signal-to-noise ratio. When combined with forward numerical simulation, they enabled a more accurate interpretation of the actual detection profiles. This study provides a practical basis for the further development and application of 3D GPR in the advanced detection of hidden geological bodies in coalfields.

Fault interpretation plays an important role in the field of coal mine safety. The development of neural network gives rise to many intelligent seismic data interpretation and processing schemes based on neural network algorithm. This study ①selected the Faster R-CNN target detection algorithm more suitable for fault recognition by comparing different deep convolutional neural network target detection algorithms. ②tested AlexNet, residual network ResNet50 and ResNet101 feature extraction networks through seismic forward modeling with various geological characteristics. It is found that ResNet101 feature extraction network has better performance in fault detection. ③constructed a fault detection model based on the preferred ResNet101 feature extraction network and Faster R-CNN target detection algorithm, and detected the actual seismic data. Results show that the object detection algorithm based on deep convolutional neural network shows satisfactory generalization ability in fault detection. It could improve the fault interpretation efficiency, and has potential in application.

In order to improve the accuracy of rock porosity test results and optimize the experimental teaching effect, this paper improves the porosity tester based on the two-chamber method of Boyle's law, and puts forward specific measures to improve the test accuracy. The coal rock of Zhaozhuang Mine in Shanxi Province, the shale and coal rock of Wannian Mine in Hebei Province, and the sandstone and coal rock of Yongcheng in He’nan Province are selected for test and analysis. The results show that the air tightness of the device, temperature change and sample drying time are the key factors affecting the accuracy of porosity test. The rock with poor pore connectivity and small porosity has higher requirements for the air tightness of the device, and the test time needs to be increased. During the experiment, the fluctuation of the ambient temperature will affect the test value, and the temperature correction is needed. The drying time of coal rock is about 48 h, and the drying time required for the quality of sandstone, shale and oil shale to achieve basic stability is shorter.

Elastic anisotropy of shales is critical to accurate constraints for rock physical models, quantitative interpretation and hydraulic fracturing. However, the causes of elastic anisotropy of shales are very complicated, and the understanding of how multiple influence factors affect the elastic anisotropy of shales is still not clear. Hence, the orthogonal experiment, as an effective multiple factors experimental method, is adopted in this study to analyze the effect of multiple factors for shale elastic anisotropy. Three factors, clay content, organic matter (OM) content and compaction stress are selected as independent variables, the orthogonal test table L₁₆(4³) with four levels for each factor is adopted. According to the designed orthogonal table, sixteen artificial shales are constructed based on the cold-pressing method, and all the dry artificial shales are measured by the ultrasonic measurements. The influence of each factor on the elastic anisotropy and the sensitivity orders of three factors are obtained using the range analysis. The orders of sensitivity for selected factors follow the sequence clay content > compaction stress > OM content for velocity anisotropy parameters. The compaction mechanism of artificial shales is also discussed by the compaction factor, which are positively correlated with the velocity anisotropy parameters. The clay platelets orientation distribution function (ODF) of samples is evaluated by a theoretical model, the ODF coefficients are significantly affected by the clay content and compaction stress, and W₂₀₀are much more sensitive to these factors than W₄₀₀. The results can provide a critical rock physics basis for quantitative interpretation and reservoir prediction of the low-maturity or maturity shale reservoir.