Abstract:This study investigates the differences in roof caving structure and crack evolution between gob-side entry retaining and traditional coal pillar mining. Taking the 52605 and 52606 working faces of Daliuta Coal Mine as the engineering background, two sets of similar-material simulation experiments were conducted to reproduce the mining processes under both conditions. The movement and fracture evolution of the overlying strata were systematically recorded and analyzed. The results show that, under gob-side entry retaining with flexible formwork concrete walls, the crack development rate at the end of primary mining is lower than that during secondary mining. In contrast, under traditional coal pillar mining, the crack evolution patterns on both sides of the coal pillar are similar. Significant differences are observed between the two mining methods in terms of crack rate, crack type, caving range, and caving angle. Specifically, for gob-side entry retaining, the crack rate of overlying strata reaches 5.0756%, the caving range extends to within 50m, and the caving angle varies from 31° to 86.9°. For coal pillar mining, the crack rate is 2.8604%, the caving range is within 40 m, and the caving angle ranges from 50°and 52°, with shear cracks dominating along the caving direction. After mining with gob-side entry retaining, the roof strata on both sides of the concrete wall remain stable without sliding, forming a hinged structural system. In contrast, in coal pillar mining, the overlying strata on both sides of the pillar tend to fail together as a whole after extraction. These structural differences lead to distinct load transfer mechanisms, resulting in significant stress concentration effects on the concrete wall in gob-side entry retaining faces.

Abstract:This study investigates the release patterns of major cations from coal and gangue collected from the Pingdingshan mining area under various conditions, including particle size and solution chemistry, with the aim of elucidating the underlying mechanisms controlling mine water mineralization. Laboratory experiments were conducted to analyze ion release characteristics and pH evolution. The results show that particle size plays a critical role in controlling ion release and solution pH. Finer particles enhance the dissolution of aluminosilicates and carbonate materials in coal-bearing strata, leading to enhanced release of key ions such as Si⁴⁺ and Al³⁺. Gangue is identified as the primary source of Si⁴⁺ and Al³⁺, whereas coal predominantly contributes Ca²⁺; moreover, the presence of gangue suppresses the release of Ca²⁺. Ion release patterns in actual mine water differ from those observed in deionized water, with Ca²⁺ and Mg²⁺ concentrations in mine water mainly governed by coal. Overall, the findings demonstrate that the interactions between coal and gangue significantly impacts the geochemical evolution of mine water, while the initial solution environment primarily regulates the dynamics of ion release and pH changes.

Abstract:This study investigates the stability of dissolution zones formed during chamber solution mining at the Pengshan Tongle glauberite mine. An integrated approach combining laboratory leaching experiments and FLAC3D numerical simulations was adopted to systematically evaluate the volumetric changes of dissolved materials and their impact on structural stability. A 3D numerical model was developed to analyze stress distribution, displacement patterns, plastic zone evolution, and surface subsidence. Experimental results reveal that dissolution shrinkage increases with higher glauberite content and decreases with larger ore particle size. Numerical analyses further demonstrate that the dissolution contraction ratio plays a critical role in controlling zone stability. After leaching, the vertical stress in the dissolved zone is reduced to approximately 3.25 MPa, compared with 13 MPa in intact pillars. In roof-connected scenarios, roof subsidence is limited to about 7 mm, with corresponding surface settlement of 7.7 mm. In contrast, non-connected conditions result in significantly larger deformations, with roof displacement reaching 37.7 mm and surface subsidence of 34.5 mm. These findings demonstrate that roof-connected leached material can provide effective structural support and maintain overall stability, whereas non-connected configurations pose a higher risk of instability. The study offers valuable insights for the safe and sustainable design of solution mining operations in evaporite deposits.

Abstract:This study aims to elucidate the effects of dissolved oxygen (DO) on aerobic granular sludge (AGS) systems treating influent with different organic matter concentrations. Two AGS reactors, i.e. R1 (low organic load) and R2 (high organic load), were operated under DO ranges of 4 mg/L to 6 mg/L and 2 mg/L to 4 mg/L to investigate differences in pollutant removal performance, microbial community structure, and functional gene profiles. The results show that after reducing the DO concentration maintained high removal efficiencies of chemical oxygen demand (COD) and total phosphorus (TP) in both reactors. Meanwhile, the rates of endogenous denitrification coupled with simultaneous nitrification increased by 17.54% and 7.05% in R1 and R2, respectively, with corresponding increases of 9.84% and 6.11% in their contribution to total nitrogen removal. Lower DO levels also induced shifts in microbial community structure, enriching functional microorganisms associated with nitrogen and phosphorus removal. In addition, the abundance of genes related to denitrification and intracellular carbon utilization increased, promoting enhanced nutrient removal performance in the AGS systems. Furthermore, DO variation exerted a more pronounced effect on the low-organic system (R1), indicating that more accurate DO control is required in such conditions for optimal operation.

Abstract:Accurate estimation of atmospheric environmental capacity is essential for optimizing industrial structure and assessing the development potential of industrial parks; however, single models often lack predictive accuracy. To address this issue, this study develops a smoke-integral multi-dimensional multi-box model based on site-specific environmental parameters and applies it to an industrial park in Chongqing. Results indicate that the calculated atmospheric environmental capacity exceeds current pollutant emissions, confirming potential for further development under existing emission controls. Comparison with the modified A-value method demonstrates minimal discrepancies, revealing the reliability and accuracy of the proposed model. Furthermore, by combining model results with current emission characteristics, targeted optimization strategies for industrial park planning and development are proposed.

Abstract:To address the low efficiency and accuracy of manual testing in railway computer interlocking systems, this study proposes a deep learning-based method for text localization and recognition in interlocking interface images. First, a text localization model based on the connectionist text proposal network (CTPN) is developed. By comparing multiple backbone networks (ResNet50, AlexNet, ZF and VGG16), VGG16 is selected as the feature extractor to enhance high-level semantic representation and improve the detection of small text regions. Second, the generalization ability and robustness of the CTPN model are improved through performance comparison with common object detection models and the incorporation of dropout. A projection-based segmentation method, combining horizontal and vertical projections, is further employed to address text adhesion issues in the interface. Finally, an improved AlexNet model is used for text recognition. Experimental results on a railway interlocking interface dataset in the TensorFlow environment show that the proposed method achieves a localization accuracy of 87.98%, a recall of 73.33%, and an F-score of 80.39%, while the recognition accuracy reaches 89%. These results demonstrate that the proposed approach can effectively locate and recognize interface text, providing reliable data support for automated routing and test result analysis in interlocking system testing.

Abstract:To address the issues of high computational complexity and large parameter size in convolutional neural network (CNN)-based image dehazing, this study proposes a lightweight dehazing network (LDNet). First, the atmospheric scattering model is reformulated to directly suppress haze noise, thereby reducing cumulative errors in intermediate variable estimation. Second, a reverse residual network module with an attention mechanism (RNAM) is designed to extract multi-scale features while emphasizing critical semantic information, effectively reducing model complexity and parameter size. Finally, a joint loss function combining L1 smoothing loss and multi-scale structure similarity (MS-SSIM) loss is used to improve reconstruction quality. The experimental results show that the proposed method outperforms existing approaches in terms of structural similarity and peak signal-to-noise ratio (PSNR) on synthetic datasets, while also achieving effective dehazing performance on real-world images. In addition, the model exhibits reduced parameter size and improved computational efficiency.

Abstract:Traditional single-frame pose estimation methods in simultaneous localization and mapping (SLAM) often suffer from cumulative errors, map misalignment, and trajectory drift due to unreliable inertial measurement unit (IMU) data and sparse point cloud features. To address these issues, this study proposes an enhanced pose estimation method based on multi-frame data fusion and optimized front-end scan matching. The proposed approach performs multi-frame fusion of LiDAR data by exploiting pose transformation relationships between consecutive frames. A weighted LiDAR-IMU fusion strategy is employed for pose prediction. During scan matching, statistical filtering is introduced to remove point cloud noise, and pose estimation is further refined through a secondary matching process. Experimental results demonstrate that, compared to mainstream conventional algorithms, the proposed method improves localization accuracy in real-world scenarios by 28.4%, 30.1%, and 65.3%, respectively, effectively reducing cumulative errors and enhancing trajectory estimation accuracy and mapping quality. This study provides a novel solution for enhancing pose estimation accuracy and mitigating cumulative errors in mobile robots mapping and self-localization tasks.

Abstract:In body area network (BAN) environments, traditional access control models face challenges such as single points of failure, rigid permission structures, and limited support for dynamic authorization. To address these challenges, this study proposes a blockchain-based dynamic trust delegation access control model for BANs. To effectively reduce storage and computational overhead, a lightweight two-layer blockchain architecture is designed, in which global policy management is maintained on the main chain, while specific service operations are processed on the subchain. In addition, a multi-smart-contract access control framework is developed to enable the automated management and execution of delegated authorization. To support dynamic permission adjustment, a trust evaluation mechanism integrating identity credibility, behavioral history, and real-time physiological context is further introduced. Experimental results show that the proposed model significantly reduces permission verification delay and emergency access latency, improve the success rate of delegation operations, and effectively reduce storage overhead. Overall, the model provides secure, efficient, and flexible access control support for resource-constrained body area network environments.

Abstract:With the continuous expansion of water resource allocation projects, accurate electricity consumption forecasting is crucial for energy conservation, cost control, and construction efficiency. Traditional forecasting methods, such as long short-term memory (LSTM) networks and Transformers, often struggle to capture both short-term and long-term dependencies in complex time-series data. To address this challenge, this paper proposes an xLSTM (extended long Short-term memory) model for multi-regional power consumption forecasting. The xLSTM model combines the short-term dependency modeling capability of sLSTM with the long-term dependency learning capacity of mLSTM, enabling effective analysis of power consumption data across multiple regions while considering temporal correlations among regions. Experimental results show that xLSTM achieve superior predictive performance, with a mean square error (MSE) of 0.0030 and a mean absolute error (MAE) of 0.035, outperforming competing models. The proposed model provides effective technical support for precise electricity demand forecasting and offers practical value for decision-making and intelligent scheduling management in large-scale water resource allocation projects.