To investigate the occurrence characteristics of the tailings deposit and slurry flow behavior during the damming process of tailings impoundments, the flume test model was used to carry out the fine-grained tailings accumulation damming test under single pipe slurry discharge conditions. The variation characteristics of beach slope, tailings particle size distribution, water level height and pore water pressure during the accumulation process of fine-grained tailings were studied. The tailings slurry flow and trajectory characteristics were analyzed, and the mechanism of tailings refinement along the deposition course was discussed. The results show that deposition slope decreased with the increase of the downstream distance. The particle size of tailings gradually becomes finer with increase of downstream distance, and the particle size is staggered with the increase of depth. The height of water level of tailings dam increases with the increase of height of dam, exhibiting a steep front slope and a gentle back slope overall. The pore water pressure rises during the discharge stage of tailings slurry and dissipates during the intermittent stage when tailings slurry stops discharging. The tailings slurry flows in a fan-shaped divergent pattern at the beginning of flume, in an irregular stream-like pattern in the middle section, and transitions from laminar flow to irregular rotational flow at the end. The slurry flow rate on dry beach surface fluctuates greatly, and slurry movement tends to be stable after reaching the tailings water surface. There is a significant difference in slurry movement characteristics before and after the interface between dry beach and tailings water.

In order to accurately locate the hidden spontaneous combustion area inside the broken coal pillar using elastic wave detection, the response characteristics and mechanism of the elastic wave velocity of coal to temperature were experimentally studied. Firstly, coal samples were pretreated by temperature programmed furnace to obtain coal samples with different oxidation degrees. Then, elastic wave and low-field NMR tests were conducted to analyze the characteristics of elastic wave and pore cracks of coal with different oxidation degrees. Finally, combined with thermogravimetric(TG) analysis and single variable briquette experiments, the elastic wave response characteristics and mechanism of coal oxidation spontaneous combustion were revealed. The results show that the P-wave and S-wave velocities of raw coal gradually decrease with increasing temperature, and the P-wave velocity decreases 2.4 times that of the S-wave velocity after the maximum mass loss temperature of TG curve. The decreasing trend of P-wave velocity shows an obvious segmental characteristic. Before the maximum mass loss temperature of the TG curve in stage Ⅰ, the P-wave velocity decreases at a rate of approximately 1.27 (m·s^-1)/℃. from the maximum mass loss temperature to the final temperature, the decline rate of P-wave velocity increases to 4.51 (m·s^-1)/℃. The porosity of coal increases from 1.57% at 30 ℃ to 9.6% at 360 ℃. In particular, after the temperature exceeds the maximum mass loss temperature, the number and aperture of cracks increase significantly, which not only lengthens the propagation path of elastic waves, but also decreases the propagation speed of elastic waves. As the propagation medium of elastic waves, the coal matrix does not decompose obviously before the maximum mass loss temperature, and the corresponding P-wave velocity of briquette remains stable at about 900 m/s. After exceeding the maximum mass loss temperature, the coal matrix decomposes rapidly, and the corresponding P-wave velocity of briquette drops to 798 m/s at 360 ℃, which proves that the coal matrix gradually transitions from a "high wave velocity" medium to a"low wave velocity" medium.

In order to evaluate the effectiveness of aerogel fire suppressants in preventing and controlling thermal runaway of lithium-ion batteries, this study selected aluminum-shell ternary lithium-ion battery cells and 18650 battery modules as experimental subjects. By employing differentiated fire suppression approaches (including application methods and operational parameters), comparative experiments on extinguishing efficacy were conducted in a systematic manner, with parallel comparisons to water-based fire suppression systems. The results show that the aerogel extinguishing agent has better fire extinguishing performance and cooling effect than water for different quantities of ternary aluminum-shell single lithium batteries and 18650 battery modules. For four ternary aluminum shell single lithium batteries, using water to extinguish fire will cause the battery temperature to rise to 492 ℃, which cannot quickly and effectively reduce battery temperature. The extinguishing speed of an aerogel extinguishing agent is 4.5 times that of water, and it can effectively prevent the battery from reburning. For 8 ternary aluminum shell single lithium batteries and two 18650 battery modules, the aerogel extinguishing agent still shows a good fire extinguishing effect, short fire extinguishing time, significant cooling effect and no reburning. Use of the aerogel extinguishing agent and fire extinguishing strategies of continuous spraying with large flow, small flow cooling after large flow spraying and spraying and flooding in containers have different fire extinguishing effects. Results show that the aerogel extinguishing agent can not only effectively extinguish initial battery fire, but also its preheating foaming mechanism and good foam stability can effectively reduce battery temperature and prevent further mixing of combustible gas and air.

In order to solve the problems of the high cost of multi-channel video data collection and long-term high-quality annotation in high-rise buildings or complex open building environments, the generation of multi-channel video data across the field of view and the automatic annotation of pedestrian images was realized. Firstly, a virtual reality scene was designed to simulate pedestrian movement and automatically obtain marker data. Secondly, unsupervised domain adaptation methods were researched to reduce the difference in feature distribution between source and target domain data, enabling the model to generalize to the target building scene. Finally, the model's generalization ability was verified. Results show that the constructed virtual reality scene effectively overcomes the difficulties of cross-visual video data collection and high-quality annotation. The unsupervised domain adaptation method increased the average first hit rate from 22.02% to 45.48%. By combining source domain style conversion, data augmentation, and target domain pseudo label generation, the first hit rate has been increased by 20%, reducing distribution bias and achieving generalization of the model in different building scenarios.

In order to solve the problem of mixed heavy metal substances in coal hindering the infiltration of dust suppression solution into coal pores and cracks, resulting in a weakened wetting effect of coal, experiments were conducted to explore the synergistic effect of chelating agents and surfactants on coal wettability. Firstly, based on the water injection test platform and the dust cutting simulation test platform, six injection methods were designed to experimentally determine the antagonistic effect between surfactants and chelating agents. Then, the effects of six injection methods on the dust production characteristics and wetting effect of coal samples were studied through cutting tests and surface wettability analysis. Finally, spray dust suppression was conducted to test the settling efficiency of pulverized coal enhanced by different injection methods. The results show that the coal sample treated by first injecting the chelating agent followed by the surfactant exhibits the lowest dust production concentration, at only 170.93 mg/m³, which is 64.49% lower than the 481.33 mg/m³ in the pure water group. Furthermore, the sequential injection protocol employing prior a chelating agent infusion followed by a surfactant delivery exhibits the most significant improvement in coal wettability, with the fastest rate of wetting-induced sedimentation of coal dust. In addition, the coal samples treated by the above method achieve a dust suppression efficiency of 91.2% in in spray dust suppression tests. Through molecular simulation and theoretical analysis, injecting a chelating agent solution to decompose heavy metals in coal fractures can improve the fluidity of the solution in coal. Subsequently, injecting a surfactant solution further enhances the wettability of coal.

The nuclear power display control system is a typical cross-screen interaction scenario in which the manipulator must accurately and rapidly monitor and respond to multiple sources of fault alarms, and the attentional capture effect plays a critical role. This study designed variables from the two dimensions of cross-screen interaction and fault alerts. In a single trial, variables such as the number of cross-screens, cross-screen orientation, the number of alerts, alert level, and alert location were set in random combinations. Eye-tracking technology and behavioral experiments were conducted to capture operators' visual attention behavior under different interaction conditions. The research reveals that an increase in the number of cross-screen displays adversely affects the attention capture effect, causing operators to frequently shift their perspectives to locate target stimuli and reducing task efficiency. As the number of alerts increases, there is a declining trend in the number and duration of fixations, indicating that excessive alert signals may lead to information overload and hinder operators' in-depth processing of individual alerts. High-level alerts effectively capture attention and facilitate rapid responses. In the future, considering the distribution of users' attention and task efficiency in cross-screen behavior scenarios, by optimizing the layout of monitoring screens and the fault alert system, improving alert design, and reducing the need for perspective transfer, the safety and reliability of alert detection task operations can be enhanced.

In order to enhance the safety and operational maintenance capabilities of tunnels, a quantitative evaluation framework and design methodology for FRP-PCM composite reinforcement in diseased tunnels was established through cross-sectional ultimate bearing capacity analysis. Numerical tests were conducted using the finite difference method to analyze the stability of FRP-PCM-reinforced diseased tunnels. Tunnel safety assessment was performed based on the sectional ultimate bearing capacity curve, commonly referred to as the M-N curve. A safety factor α of the M-N curve was proposed to enable a quantitative evaluation of the reinforcement effect. Based on this framework, a design and selection scheme for FRP was established, incorporating tunnel design specifications. The analysis reveals that when cavities exist at the crown of the tunnel, both the crown and shoulders exhibit higher risk levels according to the M-N curve. Furthermore, under conditions of poorer surrounding rock quality, greater lining concrete deterioration, and increased loosening pressure height, the reinforcing effect of the FRP-PCM method becomes more significant. For cases with an 80% degradation degree, Class V surrounding rock, and a loosened pressure height of 2D, the application of different types of FRP reinforcement reduces the safety factor of the M-N curve by 23%-32%, thereby significantly improving the structural safety of the lining. By leveraging the M-N curve's safety coefficient, an appropriate FRP type can be selected for defective lining reinforcement based on factors such as rock quality, loose pressure height, and lining conditions.

To address the perception inaccuracies of autonomous mining trucks in open-pit mines under low-visibility and low-illumination conditions—issues that may lead to obstacle collisions. This paper was proposed a multimodal fusion-based obstacle detection method to enhance detection accuracy and operational safety in complex environments. Firstly, local Feature matching at light speed (LightGlue), was employed to achieve spatial alignment between thermal infrared and visible light images, thereby avoiding spatial misalignment and geometric distortion prior to fusion. Secondly, in the modality feature extraction and fusion stage, a Dual-Modality Feature Fusion (DMFF) module was incorporated into the improved dual-branch backbone network. The extraction capability of dual-modality features was enhanced and fusion was performed through feature compression and cross-modal feature enhancement. An iterative learning method was then introduced to effectively match the complementary information between modalities, generating a fused dual-modality feature map and improving multimodal detection performance. Finally, the fused feature maps at multiple scales were input into the detection head. They were combined with bounding box regression and classification prediction for precise detection. Experimental results demonstrate that the proposed method achieves excellent obstacle detection performance in challenging scenarios with low visibility. Specifically, it achieves a mean Average Precision (mAP@0.5) of 90.8%, and an F₁-score of 0.887, outperforming existing methods in both accuracy and speed. Moreover, the proposed approach exhibits lower false positive and miss detection rates, effectively ensuring the safe navigation of autonomous mining trucks in complex operational environments.

In order to solve the multi-period and multi-batch relief delivery problem under flood disaster scenarios, a routing model was constructed for collaborative relief delivery of trucks and assault boats by comprehensively considering factors such as the dynamic changes of disaster situations, and the collaborative transportation of trucks and assault boats. The objective of the model was to minimize the total distribution time within each period. A HHA, which combines the improved ant colony algorithm(IACA) with a genetic algorithm(GA), was designed for solving this problem. Compared with the ant colony algorithm and the simulated annealing algorithm, the experimental results show that the HHA provides truck-assault boat collaborative distribution routing solutions under the dynamically changing disaster situations in a relatively short time. The HHA effectively reduces the total distribution time and improves the efficiency of relief delivery. The rationality, feasibility, and effectiveness of the HHA were demonstrated. The number of flooded areas and the number of assault boats have a significant impact on the routing of relief delivery. Emergency logistics decision-making departments should weigh the disaster situation and resource allocation of each flooded area, and dispatch a reasonable number of assault boats to ensure the efficiency of relief delivery.

In order to improve the accuracy of the risk field model in predicting the lane-changing time of the driver in the weaving area of the freeway exit ramp, firstly, the driver behavior questionnaire survey method was used to cluster the driving style and lane-changing style, and the real vehicle driving test was carried out to obtain the vehicle lane-changing trajectory data and analyzed the risk perception level of different driving styles. Then, the derived traffic environment risk was introduced on the basis of the basic risk field model, and the risk field model was constructed from three aspects: road environment, traffic environment and management control, considering the factors such as driving style and traffic control. Finally, the vehicle lane change simulation control strategy and test were designed by combining the uniform acceleration (CA) motion model and the risk field model. The results show that the motor risk perception coefficients of cautious, ordinary and aggressive driving style drivers are 0.71, 1 and 1.17 respectively, and the distance risk perception coefficients are 0.62, 1 and 1.39 respectively. The risk intensity of driving style cautious type > ordinary type > aggressive type, and the critical lane change distance of driving style cautious type > ordinary type > aggressive type. The management control risk field model predicts that the threshold time of lane change decision for cautious, ordinary and aggressive drivers is 96.2 s, 105.2 s and 114.2 s, respectively, and the error is less than 1.5% compared with the simulation lane change time of 97.4 s, 104.8 s and 113.8 s.