To prevent accidents and disasters, such as the collapse of existing roads nearby, caused by deep and large foundation pit excavation and support, it is necessary to ensure the safe operation of adjacent roads during the construction of deep and large foundation pits. Taking a deep and large foundation pit project near a city trunk road in Kunming as an example, on the basis of in-depth research at the site, the three-dimensional finite element numerical simulation software-new eXperience of GeoTechnical analysis system(MIDAS GTS NX) is used for simulation and calculation, and combined with on-site monitoring, to analyse the force and deformation characteristics of the foundation pit support structure and the deformation characteristics of the adjacent existing road under the existing excavation and support scheme. The study results indicated that the forces and deformations of the supporting piles and anchor cables were within design limits after the foundation pit excavation and support were complete. The displacement near the existing road increased with the increase of pit excavation depth, and become stable after the excavation was completed. The maximum deformation of the existing road occurred at a position 2.5 times the excavation depth from the foundation pit boundary. Furthermore, the deformation did not reach the alarm threshold for road displacement caused by foundation pit excavation. Thus, the existing support scheme can ensure the safety of both the foundation pit and the adjacent road.

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.

In order to make up for the deficiency of VR technology in safety education theory support and practical effect tests, the relevant influencing factors affecting the effect of engineering safety education were taken as the entry point. SEM was used to establish an analytical framework for the influencing factors of safety education and the cognitive experiments were carried out to test the effect. The results show that the frequency of education, the mode of education, individual initiative, fun and comfort are the main factors affecting the effectiveness of construction safety education. The hybrid mode of "VR + lecture" is the most effective educational method at present, but the use of VR technology still requires personnel assistance. The optimal frequency of education is recommended to be less than 30 days. The main factors of fun, comfort are significantly positively correlated with educational effect, but comfort is more important in the application of construction safety education with VR technology.

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 address the high rates of missed detections and false alarms, as well as the poor robustness in fatigue driving detection for open-pit mine truck drivers, a fatigue driving detection model for mine truck drivers (EBS-YOLO) based on the improved YOLOv8 is constructed to enhance the overall performance of fatigue detection. Firstly, YOLOv8 was used as the basic model for fatigue detection, and a small target detection layer was added to enhance the model's attention to small targets. Secondly, the bottleneck attention module (BAM) was used to improve the model performance to extract small target features, especially eye features. Finally, all cross-stage aggregation modules (C2f) in the backbone network were replaced with efficient multi-scale attention (EMA) modules, thereby effectively reducing model parameters and computational overhead to meet the requirements of a lightweight model. The results showed that the improved YOLOv8 model had a great detection effect with the accuracy, recall rate, and average detection accuracy reaching 93.6%, 93.9%, and 96.5%, respectively, and the memory size of the model was only 4.9 MB. Compared with the YOLOv8 model, the improved model can quickly and accurately identify the fatigue state of mining truck drivers, meet real-time requirements, and effectively prevent fatigue-driving accidents.

To avoid worsening the consequences of oil and gas pipeline accidents due to emergency failures，the causes of emergency failure in 27 accidents at home and abroad were analyzed using the HFACS model. Based on the results of grounded theory (GT) statistical coding analysis，a classification model of failure causes of emergency response in oil and gas pipeline accidents was proposed. SNA was used to develop the relationship network of the causes of emergency failures in oil and gas pipeline accidents. The core-periphery，centrality，and association direction index analyses were used to identify core factors and factors with high association and strong mediating roles in the classification model of the causes of emergency failures in oil and gas pipeline accidents. The results indicated that the classification model of emergency failure causes in oil and gas pipeline accidents was divided into five levels: government and emergency department factors，operator organizational factors，operator unsafe supervision，preconditions for unsafe behavior of on-site personnel，and unsafe behavior of on-site personnel. The emergency failure causes were further divided into 16 bottom-level factors，among which there were 9 core factors: inadequate safety supervision by government and emergency departments，ineffective emergency rescue，regulations defects，insufficient supervision by pipeline operators，technical environment，and skill errors. Skill errors，regulations or procedure defects，technical environment，and insufficient supervision by operators were highly associated factors. Moreover，pipeline operators' regulation defects，procedure defects，technical environment，insufficient supervision，improper resource management，and decision-making errors were strong mediating factors.

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 explore the effect of cognitive load on workers' hazard identification behavior，a cognitive experiment based on eye-tracking technology for construction site hazard identification was designed firstly. In this experiment，a N-digit task was introduced to increase the cognitive load. Secondly，the gaze and glance data were collected to analyze the static attention allocation，and the scanning path was processed to extract the dynamic transfer characteristics of attention. Finally，three parameters of variance review probability (RP)，transition probability (TP) and switching probability (SP) were selected as the quantitative parameter values to classify the scanning patterns of hazard identification，which explored the influence of cognitive load on hidden hazard identification from the perspective of visual behavior performance. The results show that the level of cognitive load negatively affects hazard identification performance. The subjects with high cognitive load show longer first fixation time，fewer fixation counts and saccade counts，and there is no significant difference in fixation percentage and mean fixation duration. Additionally，based on the attention characteristics，three scanning patterns are identified: sequential inspection，repeated comparison and random discovery. With the improvement of cognitive load level，subjects will pay more attention to identifying single hazard but neglect others during sequential inspection，and reduce the attention in the hazard area but still keep the fixation point quickly and frequently switching during repeated comparison，while the number and time of inspection of hazard areas are reducing simultaneously during random discovery.

To clarify the essential characteristics and differences between inherent safety, behavior-based safety, process safety, and functional safety and to promote a virtuous cycle of high-quality development and high-level safety, this study employed literature review and comparative analysis methods to explore their basic connotations and evolution processes, interrelationships, realistic challenges, and development paths based on the safety management paradigm shift. The results indicate that inherent safety is an idealized form of safety. Behavior-based safety is an interdisciplinary field that involves the theories and methods of safety science and behavioral science. Process safety protects humans, machines, and the environment through systematic approaches from a full life cycle perspective. Functional safety aims at preventing unacceptable risks caused by functional failures of systems. The four types of safety, led by inherent safety, involve a gradual progression from concepts to practice. These types of safety share a unified internal structure encompassing the elements of humans, machines, environment and management. The current representative standards cover various industry sectors and focus on accident prevention. In the future, the synergistic effect of the four in safety governance should be fully utilized. By using artificial intelligence technology to empower the new engine of safety production, the four should be continuously improved in specific practices tailored to local conditions.

Cultivating professional master's students is essential to addressing the shortage of high-level applied talents. To meet the industry demand for safety engineering professionals, this study analyzes challenges in China's training processes based on domestic and international practices. It introduces the "professional group + action learning method" model, alongside reforms in curriculum, teaching methods, training bases, faculty, and evaluation standards, using China University of Mining and Technology-Beijing as a case study. Data from the 2023 cohort validate the model's effectiveness in improving graduate quality, enhancing competencies, and addressing traditional education shortcomings, proving its feasibility and reference value.