To improve the interaction ability of traffic vehicles in the cut-in scenario，a method for constructing a vehicle hazardous cut-in strategy based on deep reinforcement learning was proposed. Firstly，a simulated environment was built based on scalable multi-agent reinforcement learning training school(SMARTS) simulation platform. Then，twin delayed deep deterministic policy gradients (TD3) algorithm was adopted to train an agent to cut in a randomly chosen target vehicle hazardously. The algorithm was compared with proximal policy optimization (PPO) and deep deterministic policy gradient (DDPG) algorithms. The trained model was tested in seven different scenarios with varying traffic densities. Finally，a multi-agent testing environment was built，and the trained model was applied to validate intelligent driving strategies. The results show that the success rate of hazardous cut-ins reaches 80.35% in model training with TD3 algorithm，outperforming both comparative methods. In model testing，except for the 2 700 vehicle/h test scenario，the model achieves a hazardous cut-in success rate of over 80% in the other three test scenarios that were not used in training，demonstrating good generalization ability. Meanwhile，the time to collision values between the ego vehicle and the target vehicle at the moment of lane changes are concentrated within the range of 0 to 6 seconds，with 95% falling within this bracket. The proportions of time to collision values in the intervals of (0，2]，(2，4]，(4，6]s are 60%，30%，and 5% respectively，covering test conditions with different collision risk. In the validation of intelligent driving strategies，the traffic vehicle controlled by the trained model can actively perform cut-ins in front of the test vehicles，exposing it to the risk of a rear-end collision and helping in identifying safety vulnerabilities in intelligent driving strategies.

To improve the safety risk prediction accuracy of injection-production string, a joint simulation analysis method was used to perform finite element verification of the fatigue crack propagation process of straight-notch compact tensile specimens. Then, a finite element model of fatigue crack propagation of injection-production string with external surface cracks in gas storage was proposed. Moreover, the injection-production strings' fatigue crack propagation behavior was analyzed under alternating pressure loads. The results indicated that the specimen simulations were consistent with the fatigue crack test results, indicating that the joint simulation method had high accuracy in fatigue crack propagation analysis. During the fatigue crack propagation of the injection-production string, the larger the initial crack's length-to-depth ratio or the higher the stress ratio, the faster the crack propagation rate with a critical minimum crack length of 4 mm. Under the same number of pressure load cycles, the larger the initial crack's circumferential angle, the longer the surface crack propagates, and the critical minimum circumferential angle was 45°. The crack consistently propagated along the axis of the pipe string regardless of the initial crack's circumferential angle, and the unstable propagation length was 52 mm. During alternating injection and production in gas storage, the amplitude of the alternating pressure load should be properly controlled to avoid the initiation and propagation of fatigue cracks.

In order to study the effect of different metal doping and precipitation temperatures on the catalytic oxidation of CO by Cu-Mn type catalyst, the co-precipitation method was used to prepare Cu-Mn type CO catalyst, and the catalytic oxidation of CO by Cu-Mn type CO catalyst under different metal doping and precipitation temperatures was tested and analyzed. The pore characteristics and surface crystal structure of the catalyst were obtained by automatic physical adsorption analyzer and X-ray diffraction (XRD). The reaction process of catalytic oxidation of CO was revealed by in-situ diffuse reflection infrared spectroscopy, and the potential application of the catalyst in coal mines was introduced. The results show that during the test time (within 80 s), with the increase of reaction time, the volume fraction of CO gradually decreased, slowly increased and then tended to be flat, and the amount of reactive CO substance gradually increased. The better the catalytic oxidation of CO, the larger the specific surface area, the smaller the average pore size and the larger the total pore volume. When the doped metals are Sn, Fe and Ce, the catalytic oxidation characteristics of the three catalysts are as follows: CuMnO_x-Ce>CuMnO_x-Sn>CuMnO_x-Fe, the amount of CO involved in the reaction was 0.015 3, 0.009 3 and 0.020 3 mol, and the removal efficiency of CO was 61%, 47% and 77%, respectively. When the precipitation temperature is 70 ℃, the number of crystal nuclei of the catalyst is significantly higher than that of the precipitation temperature is 60 and 80 ℃. When the precipitation temperature is 60, 70 and 80 ℃ respectively, the catalytic oxidation characteristics of the three catalysts are as follows: CuMnO_x-Ce-70>CuMnO_x-Ce-80>CuMnO_x-Ce-60, the amount of CO involved in the reaction was 0.019 45, 0.020 3 and 0.019 8 mol, and the elimination rates of CO were 74%, 77% and 75%, respectively. Abundant surface oxygen vacancy is the key factor to improve the performance of CO oxidation reaction and catalytic oxidation. The presence of CeO₂ contributes to the formation, oxygen activation and migration of carbon-containing species.

To deeply understand the competitive adsorption characteristics of O₂/CH₄ gas mixtures on the coal surface, the pore distribution and chemical structure of the coal surface were analyzed using the Grand Canonical Monte Carlo(GCMC) method. Molecular simulation was used to investigate adsorption capacity and heat variations during the competitive adsorption of O₂/CH₄ gas mixtures under varying concentrations, pressures, and temperatures. Moreover, the competitive adsorption mechanism of O₂/CH₄ gas mixtures on the coal surface was revealed by adsorption selectivity. The results indicated that the isothermal adsorption curves were identical when the concentration ratio of O₂ to CH₄ was 7/3. The integrated adsorption heat under different conditions was approximately 19.5-20.0 J/g for O₂ and 24.0-24.8 J/g for CH₄. When the concentration ratio of O₂ to CH₄ ranged between 7/3 and 2/8, the integrated adsorption heat of O₂ was higher than that of CH₄. The increased temperature decreased the adsorption capacity of both gases but increased the adsorption selectivity of O₂ to CH₄. Moreover, the effect of gas pressure on competitive adsorption weakened with pressure increase. CH₄ adsorbed faster than O₂ under low gas pressures, and CH₄ reached saturation adsorption earlier than O₂ when the total pressure increased. When the O₂ concentration in the gas mixtures increased, the adsorption selectivity of O₂ for CH₄ decreased. However, CH₄ consistently showed higher adsorption competitiveness than O₂.

In order to investigate the threshold conditions for pedestrian destabilization, controlled experiments were conducted in three different scenarios, including "two-person collision experiment", "multiple-person collision experiment" and "high-density multiple-person collision experiment". A pedestrian flow model was constructed, considering collision pressure, speed, density and volume. Additionally, experimental videos and collision pressure data were collected by using an unmanned aerial vehicle(UAV)and pressure sensing equipment, respectively. The changing rules of pedestrian trajectory, pedestrian flow characteristics, and collision pressure were obtained by data analysis. On this basis, the segmented Hermite triple interpolation was used to investigate the influence of speed and density on collision pressure. Finally, based on the Van Aerde model, a four-dimensional model of "pressure-speed-density-volume" was constructed by introducing the parameters of collision pressure to judge the threshold conditions of pedestrian flow instability. The results show that pedestrian flow begins at 2.46 persons/m² and reaches complete instability at 3.59 persons/m². During the instability process, the pedestrian may be subjected to the collision pressure ranging from 187.32 N to 258.11 N. The results provide references for the control of the pedestrian flow and the improvement of pedestrian traffic safety.

In order to solve the problems of ambiguity and randomness in the process of fire safety resilience assessment of subway stations, and then effectively improve the level of fire safety resilience of subway stations, the fire safety resilience assessment model for subway stations based on WSR- extension cloud theory was constructed. First, based on WSR methodology, the factors affecting the fire safety resilience of subway stations were analyzed around "physical", "matter", and "human". Combined with the characterization of the resilience absorption, resistance, recovery and adaptive ability, the fire safety resilience assessment index system of subway stations was established in five aspects, namely, equipment factors, environmental factors, organization and management, material and technology, and personnel factors. Second, based on the blind number theory to construct the blind number matrix and calculate the comprehensive score of qualitative indexes, the fire safety resilience level of subway stations is derived by using the theory of extension cloud theory. Finally, a station of the Qingdao subway was used as an example to carry out the example analysis. The results show that the subway station fire safety resilience level of Ⅳ, in the higher resilience level, the credibility factor = 0.003 4 <0.01, indicating that the assessment results have a high degree of credibility. The WSR-extension cloud theory assessment model can provide a theoretical basis for the fire safety resilience assessment of subway stations.

In order to improve the evacuation efficiency of the teaching building, the classroom structure was optimized through control experiments and numerical simulations to enhance evacuation efficiency. Emergency evacuation tests were used to obtain the movement characteristics of students aged 6-7 years old. And Pathfinder simulation software was used to study the impact of desk layout, classroom door position, and exit position on evacuation. The results indicate that for a single classroom, although shortening the pre-action time can reduce the overall evacuation time, it cannot improve the congestion caused by the building structure. Appropriate evacuation routes and desk layouts can significantly reduce evacuation time. For buildings with classrooms on one side of the corridor, increasing the width of the corridor and exit is the most effective way to improve evacuation efficiency. For buildings with classrooms on both sides of the corridor, the structure of the evacuation corridor inside the building, including the number of corridors and the intersections inside the corridors, is the most important factor affecting evacuation time. Therefore, it is recommended to develop optimization plans for classroom evacuation structure from different aspects.

To reduce the risk of rear-end collisions in highway tunnels, a tunnel reflective strip space model was proposed based on the tunnel reflective strip characteristics with depth perception information. The effects of spacing between tunnel reflective strips on vehicle distance maintenance were examined through driving simulation tests. Braking headway, minimum headway, following headway, and minimum collision time were selected as evaluation indicators. The results indicated that when the leading vehicle traveled at speeds between 40-80 km/h and the following vehicle approached at a speed 20 km/h higher than the leading vehicle, if the following driver received 3-4 visual stimulations from the reflective strips before the distance between the leading and following vehicles was less than the minimum safe distance, the braking headway, minimum headway, and the time headway under stable car-following conditions was improved by 27.6%-56.6%, 54.2%-60.3%, and 20.1%-31.6%, respectively. Furthermore, the minimum collision time was increased by 34.7%-60.5% once the leading vehicle braked urgently, reducing the probability of a rear-end collision. Therefore, tunnel reflective strips enhance drivers' perception of speed and distance perception ability, reducing the risk of rear-end collisions.

In order to reduce the losses caused by landslide disasters, taking Changde city of Hunan province as an example, the fractal theory and information method were applied to evaluate the regional landslide susceptibility based on field investigation and historical landslide data. The sensitivity of influence factors was quantitatively studied by fractal theory. The information values of each secondary impact factor were obtained by using the information method, and the comprehensive information values were obtained by combining the fractal dimension value and the information value. Based on the values, the susceptibility zoning of the study area was carried out. The results show that the slope, engineering geological rock group, elevation and vegetation are the influencing factors that have a second-order cumulative and fractal distribution with the landslide, while other influencing factors have a first-order cumulative and fractal distribution with the landslide. The areas of very low, low, medium, high and very high susceptibility areas respectively account for 5.24%, 8.84%, 35.06%, 39.21% and 11.65% respectively. Annual rainfall greater than 1 600 mm, slope of 20-30° and elevation of 900-1100m are important factors.

In the early stages of a major disaster, where demand at disaster sites is uncertain, roads are damaged in affected areas, and the fairness and timeliness of rescue operations must be considered, the SNS algorithm is applied to solve the emergency supply distribution model to achieve rapid and effective distribution of emergency supplies. First, an emergency supply distribution model was constructed, with the objective of minimizing the total cost of emergency rescue and the evaluation of humanitarian aid under the background of fuzzy demand and damaged roads. Then, the SNS algorithm was introduced to solve the model, and an improved SNS(ISNS) algorithm with a reinforcement learning rate strategy was proposed. Finally, taking the 2022 Luding Earthquake in Sichuan as an example, the SNS algorithm, ISNS algorithm, discrete particle swarm optimization, genetic algorithm, and simulated annealing algorithm were applied to solve this case, respectively. The results indicate that the ISNS algorithm demonstrates stability. Compared with other algorithms, the total cost of emergency rescue is reduced by at least 6 410 yuan, and the evaluation of the humanitarian aid evaluation target is improved by at least 50.6%, highlighting the superiority of the ISNS algorithm. The ISNS algorithm is beneficial for solving emergency supply distribution problems.