To ensure the safety of personnel and property within the storage environment，the traditional YOLOv11 object detection algorithm was improved，and a method and model to identify unsafe behaviors of personnel in the complex environment of tobacco warehouses were proposed. First，a statistical analysis of common unsafe behavior types in tobacco storage was conducted，and the classification of unsafe behaviors of warehouse personnel was explored，including item-related，action-related，and area-related unsafe behaviors. Second，based on the characteristics of unsafe behaviors of warehouse personnel，a dataset augmentation and denoising preprocessing approach was proposed to enhance fine-grained feature extraction，and introduced to improve the saliency mapping of personnel behaviors. Then，the YOLOv11 algorithm was improved through functional enhancement modules and K-means++ anchor box optimization，and a fast detection method for unsafe behaviors of tobacco warehouse personnel was proposed. Finally，the proposed method's effectiveness was validated by comparing with self-built datasets and the open Microsoft COCO dataset. The results show that the method can quickly and effectively identify unsafe behaviors of warehouse personnel，with a significant improvement in recognition accuracy compared to traditional methods(accuracy rate is 94.91% and 88.69% respectively).

In order to enrich the safety & security management theories and innovate the safety & security management paradigm，the research work of parallel safety & security management was carried out based on the parallel system theory and safety & security management theory. Firstly，starting from the parallel system theory and combining with the knowledge of modern safety & security management，the concept of parallel safety & security management was explored and the parallel safety & security management conceptual model was proposed. On this basis，the parallel safety & security management model was constructed and explained. Finally，taking urban safety & security as an example，this paper analyzed the application of the parallel safety & security management model. The research shows that parallel safety & security management is a new paradigm for safety & security management through realizing the virtual-real interaction between artificial safety & security systems and real safety & security systems to achieve multiple objectives such as solving complex safety & security problems，effective implementation of safety & security solutions and efficient training of personnel. Its model has the characteristics of breaking the constraints of real-world conditions，flexibility and scalability，high performance distributed computing，redundancy and fault tolerance，simulation prediction and continuous improvement，task scheduling and load balancing，which can provide a model reference for safety & security management of complex systems，such as urban safety & security management.

In order to improve the safety production level of coal mining enterprises and ensure a reasonable proportion of safety investment，the CPT method was introduced to establish a dynamic safety investment decision-making model that combined the criteria importance through intercriteria correlation(CRITIC) method improved by the coefficient of variation method and CPT. From the perspective of safety input and output，an evaluation index system was constructed，and the weights of each index were calculated by the improved CRITIC method. Combined with the CPT，the static decision-making ranking results of various safety investment schemes in coal mining enterprises were calculated，and on this basis，the results were dynamically optimized to obtain the final dynamic decision-making ranking results of safety investment. The study shows that the constructed model is reasonable and reliable with high sensitivity compared with VlseKriterijumska Optimizacija I Kompromisno Resenje (VIKOR) method. The model considers the influence of the decision maker's subjective risk preference on the actual decision-making results，and the time factor can reflect the dynamic accumulation level during the research period，so the decision-making results are more scientific and reasonable.

In order to investigate the leakage and dispersion patterns of gas-phase CO₂ transport pipelines with different hole sizes，outdoor CO₂ leakage test with apertures of 50 mm，100 mm and 233 mm were carried out based on industrial pipelines. Firstly，according to the pressure drop experiment，the variation law of the pressure in the pipe with the leakage time was analyzed. Then，based on the established theoretical model，the leakage flow and pressure drop changes were predicted and verified by experiments. Finally，the variation of CO₂ volume fraction at different positions of the leakage port was studied. The result shows that the pressure drop rate in the first stage of leakage is the largest. Among them，the pressure drop rate of the three leakage hole sizes is 93.4-1967.5 kPa/s. It takes 14.2-149.8 s for the pressure in the pipe with three leakage hole sizes to drop to the ambient pressure. With the increase of the leakage hole size，the pressure drop process tends to be a cliff type drop. According to the established theoretical model，the relative error between the calculated leakage quality and the experimental actual release is 0.25%-4.54%. The pressure drop curve obtained based on the predictive model is generally slightly lower than the experimental pressure drop curve. However，the variation trend and duration of pressure drop are very close to the experimental results. It shows that the established theoretical method for predicting leakage quality and pressure drop is reliable. In the range of 5-17 m from the leakage port，the peak volume fraction increased significantly with the increase of the leakage hole size. The peak volume fraction of different leakage hole size outside the range of 17 m approaches the same. The dispersion distance of 5% volume fraction of gaseous CO₂ leakage is 26 m.

In order to further scientifically prevent and control coal mine gas accidents and systematically analyze the risk factors and coupling relationships of coal mine gas accidents in my country，an association rule mining model based on the PSO algorithm using Python software was established and verified. The risk factors of coal mine gas accidents were classified in combination with the HFACS accident risk model，and the constructed PSO-FP(Freguent Pattern)-growth algorithm was used to mine association rules for coal mine gas accident investigation reports. The results show that the PSO-FP-growth algorithm has better running speed and association rule effect than the PSO-Apriori algorithm. According to the visualization of association rules of gas accident risk factors and high-support association factors，the main risk factors for coal mine gas accidents in my country are defects in safety supervision and management of coal mine enterprises，inadequate gas prevention and control technology，weak safety awareness of employees，and inadequate management awareness and technology of on-site managers.

In order to assess and monitor flight risks in real-time，clustering analysis was utilized to explore the abnormal patterns embedded in QAR data，and the influencing factors of abnormal flight patterns of civil aircraft were analyzed. The Euclidean distance was employed to characterize the similarity between samples of QAR parameters，establishing an abnormal flight pattern recognition model based on K-means to define the deviation degree of abnormal patterns. By considering the number of fatal accidents and the proportion of deaths in global commercial jet accidents，in conjunction with the deviation degree of abnormal patterns，the duration of abnormal patterns，flight phases，the likelihood of unexpected safety events，and the severity of consequences following unexpected safety events，a quantified assessment method for civil aviation flight risks based on QAR data was proposed. The feasibility of abnormal flight pattern recognition and risk quantification models for civil aircraft was validated through the practical QAR data of a certain airline. The results indicate that abnormal patterns are more prevalent during the cruising phase and critical moments at the transitions between flight phases. Significant differences are observed in the distribution of abnormal flight patterns and risks across different flights and flight phases. The average total risk value for flights is 166.94，with outliers exceeding 386.97. The abnormal flight risk during the takeoff roll phase is relatively low，with an average of 5.95，while the risk during the cruising phase is relatively high，with an average of 93.46.

To enhance the flame retardancy and smoke suppression performance of EP-CE composites，a novel flame retardant was synthesized using DOPO and BPS as raw materials. The flame-retardant DOPO-BPS/OSEP/EP-CE composites were fabricated by the solution casting method，where DOPO-BPS and OSEP were jointly introduced into EP-CE matrix. Based on Fourier transform infrared spectroscopy (FTIR) analysis，nuclear magnetic resonance (NMR) analysis，thermogravimetric analysis (TGA)，limiting oxygen index (LOI) test，UL-94 vertical burning test，and cone calorimeter analysis，the effects of different ratios of DOPO-BPS and OSEP on the flame retardancy and smoke suppression properties of EP-CE composite materials were studied. The results show that the phosphaphenanthrene ring structure of DOPO-BPS could form a stable carbon layer and block the transfer of heat and oxygen. The synergistic effect with OSEP significantly improves the flame retardant effect and smoke suppression performance. When the flame retardant DOPO-BPS and OSEP are added to EP-CE at a mass fraction of 9∶1，the flame retardant performance and smoke suppression performance are the best，the flame retardant performance reaches V-0 level，and the LOI increases to 31.4%. Compared with EP-CE，the peak heat release rate，total heat release and total smoke production decrease by 29.7%，32.2% and 32.7%，respectively.

To investigate the impact of different injection source gases on the adsorption and diffusion behavior of methane (CH₄) in coal，three types of gases were selected: hot gas power generation exhaust (heat injection，multi-component)，carbon dioxide (CO₂) at room temperature (strong adsorption，single component)，and nitrogen (N₂) at room temperature (weak adsorption，single component). Using giant canonical Monte Carlo (GCMM) and molecular dynamics (MD) methods，these gases were mixed with CH₄ and injected into coal to analyze the adsorption conditions. Based on a fixed amount of CH₄，the changes in diffusion behavior were analyzed after injecting each of the three gases. The results show that with the increase of gas injection ratio，the reduction of CH₄ adsorption capacity under CO₂ injection condition is gradually greater than that under thermal power generation tail gas condition，showing better inhibition performance than thermal power generation tail gas. In contrast，although the adsorption capacity of CH₄ decreases after N₂ injection，it is always greater than the previous two. In terms of diffusion，with the increase of gas injection ratio，the diffusion coefficient increases first and then decreases，and the coefficient is always larger than before gas injection，and the displacement gas mainly promotes CH₄ diffusion. Under N₂ injection，the diffusion coefficient of CH₄ is the highest and the decrease is the smallest，and the promoting effect is the most obvious. Under the condition of CO₂ injection，the diffusion coefficient of CH₄ decreases the most and the promoting effect is the weakest. Therefore，the selection of hot tail gas from gas-fired power generation for CH₄ displacement is more cost-effective.

To explore human error in complex emergency rescue scenarios for hazardous chemical accidents in chemical industry parks and improve human reliability in emergency rescue actions，a comprehensive analysis method was established to quantitatively evaluate the human reliability of emergency rescue for hazardous chemical accidents. Firstly，based on the laws，regulations and standards related to the emergency rescue of hazardous chemical accidents，20 emergency behaviors in chemical industry parks were summarized and extracted. Secondly，cognitive reliability and error analysis method(CREAM) was introduced to determine the probability of human error. Analytic hierarchy process(AHP) and entropy weight method were combined to quantify the severity of errors in emergency behaviors. Finally，from the perspectives of possibility and severity，the weak aspects of emergency behaviors in chemical industry parks were explored，and the strategies for enhancing emergency rescue capabilities in chemical industry parks were discussed. The practical application of the method was verified with the example of N chemical industry park. The results show that 20 emergency behaviors were divided into 3 clusters. There are 4 emergency behaviors identified that needed to be prioritized for improvement in the petrochemical zone: risk assessment，fire-fighting，initial disposal of enterprises and rescue of people in distress. For these emergency behaviors，it was proposed that N chemical industry park should focus on optimizing the accident information transmission mechanism，improving decision-making and command effectiveness，and strengthening rapid response and disaster identification capabilities，so as to provide countermeasures and suggestions for improving its emergency rescue capability.

In order to effectively control low-frequency vibrations in engineering practice，the structure and the formation mechanism of the band gap were analyzed，and the finite element method was employed to investigate the influencing factor of the band gap in combined periodic cross-isolation trenches. Subsequently，this study analyzed vibration isolation performance using the model experiment. The results show that the band gap characteristics of the combined periodic cross-isolation trenches are mainly influenced by the periodic constant，the depth of the channel，the elastic modulus of soil and the filling medium. With the increase of the period constant，combined periodic cross-isolation trenches are more likely to obtain the low-frequency band gaps，but the bandwidth narrows. By increasing the depth，the low-frequency band gap with a larger bandwidth can be obtained，and the number of band gaps increases. With the increase of elastic modulus，the boundary frequency of the band gap migrates to high frequency synchronously，and a more low-frequency and wider band gap can be obtained by filling the medium. The experimental results show that the maximum attenuation amplitude of acceleration in the band gap range reaches 0.001 24 m/s²，the average amplitude attenuation ratio in the band gap is 0.206 2，and the maximum attenuation degree reaches 98.4%. The combined periodic cross-isolation trenches have remarkable vibration isolation performance.