Mining strip coal pillars left due to strip mining is important to resource-exhausted coal mines in eastern China. Backfilling mining is an effective means to mine strip coal pillars, which could decrease high stress and high energy release. However, materials with super-high water content were not widely applied in deeply isolated coal pillar mining due to lower strength characteristics and durability. In the paper, taking panel c8301 as engineering background, theoretical analysis, numerical simulation, and field monitoring were used to study the feasibility and application effect of the super-high-water backfilling technology in deep isolated coal pillar mining. The results showed that: (1) Panel c8301 is a strip-filling working face with insufficient mining on both sides, and the mining of the panel will trigger the rebalancing of the overlying rock structure on both sides, which increases the rock burst risk. (2) Simulation results indicate that the super-high-water filling method, in comparison to the traditional caving method, significantly reduces peak stress and elastic energy from 112.3 MPa to 4.6 × 10⁶ J to 79.2 MPa and 2.23 × 10⁶ J, respectively. This represents reductions of 29.6 % and 51.5 %, effectively mitigating the impact of mining activities on overburden movement. (3) On-site measurement data confirmed that the measured equivalent mining height was 1.25 m. The total number of microseisms and the amount of released energy decreased significantly, More specifically, three big energy release instances (>1.8 × 105 J) were recorded during the first roof weighting stage and panel in square meters. Super-high-water filling technology has achieved remarkable results in the mining of strip coal pillars and has significant application prospects.

The safety valves of powered supports control the maximum working resistance, and their statuses must be known to ensure the safety of both the support and the overlying strata. However, the inspection of powered support valves involves manual or semiautomated operations, the costs of which are high. In this study, an extreme point extraction method was developed for the determination of the characteristic parameters of safety valves using roof pressure data, and a safety valve state monitoring module was constructed. Using the longwall face of 0116³06 with top coal caving in the Mindong Mine as an example, the characteristic parameters of the safety valves were extracted, including the peak, reseating, and blowdown pressures, as well as the recovery and unloading durations. The results of the field tests showed the following: (1) The amplitude threshold method based on extreme points can be used to accurately extract characteristic parameters, and the distribution of the characteristic parameters of the safety valves follows either a Gaussian or an exponential distribution. (2) The mining pressure analysis results, derived from the characteristic parameters, closely align with the in situ mining pressure observations. This method can be used for the online monitoring of safety valve conditions, increasing the operational efficiency and quality of safety valve inspections.

The precise detection of stratum interfaces holds significant importance in geological discontinuity recognition and roadway support optimization. In this study, the model for locating rock interfaces through change point detection was proposed, and a drilling test on composite strength mortar specimens was conducted. With the logistic function and the particle swarm optimization algorithm, the drilling specific energy was modulated to detect the stratum interface. The results indicate that the drilling specific energy after the modulation of the logistic function showed a good anti-interference quality under stable drilling and sensitivity under interface drilling, and its average recognition error was 2.83 mm, which was lower than the error of 6.56 mm before modulation. The particle swarm optimization algorithm facilitated the adaptive matching of drive parameters to drilling data features, yielding a substantial 50.88% decrease in the recognition error rate. This study contributes to enhancing the perception accuracy of stratum interfaces and eliminating the potential danger of roof collapse.

Environmental perception is crucial for achieving autonomous driving of auxiliary haulage vehicles in underground coal mines. The complex underground environment and working conditions, such as dust pollution, uneven lighting, and sensor data abnormalities, pose challenges to multimodal fusion perception. These challenges include: (1) the lack of a reasonable and effective method for evaluating the reliability of different modality data; (2) the absence of in-depth fusion methods for different modality data that can handle sensor failures; and (3) the lack of a multimodal dataset for underground coal mines to support model training. To address these issues, this paper proposes a coal mine underground BEV multiscale-enhanced fusion perception model based on dynamic weight adjustment. First, camera and LiDAR modality data are uniformly mapped into BEV space to achieve multimodal feature alignment. Then, a Mixture of Experts-Fuzzy Logic Inference Module (MoE-FLIM) is designed to infer weights for different modality data based on BEV feature dimensions. Next, a Pyramid Multiscale Feature Enhancement and Fusion Module (PMS-FFEM) is introduced to ensure the model’s perception performance in the event of sensor data abnormalities. Lastly, a multimodal dataset for underground coal mines is constructed to provide support for model training and testing in real-world scenarios. Experimental results show that the proposed method demonstrates good accuracy and stability in object-detection tasks in coal mine underground environments, maintaining high detection performance, especially in typical complex scenes such as low light and dust fog.

Measurement while drilling technology (MWD) has emerged as a pivotal approach for geological exploration. However, the accuracy of existing geological recognition models remains limited, primarily due to data fluctuations that result in high overlap rates and reduced reliability of drilling parameters. This study takes torque data as an example and analyzes the frequency distribution laws of torque responses across rock with varying strengths. A quantitative model of the frequency distribution characteristic interval is established, and a rock information prediction approach based on frequency distribution characteristics is proposed. The results indicate that torque frequency distributions for homogeneous rock exhibit a unimodal pattern, whereas those for composite rocks display multimodal characteristics. The boundaries of the frequency distribution characteristic intervals are mathematically defined as CIS = T_p|_{(dF/dT) = 0} ± σ and CIM = x_li ± 0.5∆x_i. The strength prediction model constructed using torque within the characteristic interval achieves an average accuracy of 85.3%. Furthermore, the frequency of torque within the characteristic interval enables the estimation of rock stratum thickness. This research contributes to enhancing the accuracy of rock information identification.