In order to improve the effectiveness of oil and gas pipeline accident prevention strategies，a classification model for the causes of oil and gas pipeline accidents was developed，and social network analysis was applied to the classification model. Firstly，the STAMP model and HFACS model were combined to get the control structure of oil and gas pipeline accident prevention，and then the causes of 35 oil and gas pipeline accidents at home and abroad were analyzed according to the control structure. The analysis results were coded using grounded theory to get the classification model for the causes of oil and gas pipeline accidents. Social network analysis methods were applied to construct a relationship network of factors related to oil and gas pipeline accidents，and core edge analysis，centrality analysis，and correlation direction index analysis were used to identify the core factors and factors with high correlation and strong influence in the oil and gas pipeline accident classification model. The research results show that the classification model for the causes of oil and gas pipeline accidents included 6 levels and 22 bottom cause factors，which are government and regulatory factors，third-party factors，operator organizational factors，operator unsafe supervision and the prerequisites for unsafe behavior of on-site personnel. Among the causal factors，the internal factors of the government and regulatory authorities，organizational factors of operators，unsafe supervision of operators，and third-party factors are core factors. System flaws，insufficient supervision，improper operation plans，third-party sabotage behavior，pipeline and weld defects，construction/repair/accessory issues，and skill errors are factors with high correlation and strong influence.

To have a deep understanding of the causal relationship between crane drivers' situation awareness information and behavior，the situation awareness and behavior response model of crane drivers was proposed by combining the ENDSLEY situation awareness model with the DEMATEL-AISM method. Firstly，the situation awareness theory was used to analyze the driver's behavioral response process and obtain information factors during the crane operation task. Secondly，the DEMATEL method was used to quantitatively analyze the association between the factors and determine a comprehensive influence matrix. Moreover，the attributes and characteristic values of the factors were analyzed to identify the key factors. Finally，a stable hierarchical structure of cause-effect attributes obtained by the AISM was used to propose the situation awareness and behavior response model of crane drivers. The results revealed that a five-layer information model consisted of 22 elements and influence relationships such as key elements of trajectory prediction and planning，and collision avoidance. Furthermore，the proposed model clarified the attribute characteristics，influence relationships，and influence degree among the information elements，offering a deep understanding of the crane driver's situation awareness and behavior response.

In order to address the challenges associated with characterizing the scenarios of stampede accidents and facilitating comprehension of these scenarios among decision-makers，a method for constructing and combining scenarios of large-scale event stampede accidents was proposed. Firstly，the scene elements of stampede accidents were extracted in large-scale events from the four factors that affect the formation of large-scale activities: people，venue，management，and environment，and a formal expression method for "state" and "trends" of large-scale activities research was established. Secondly，based on Markov model，a deduction description and calculation method for the transformation of situational "state-trends" was provided. Finally，an example analysis was conducted using Shanghai Bund accident. The findings of empirical analyses indicate that deductive results are largely aligned with the actual development process of the 2014 Shanghai stampede. This evidence substantiates the scientific rigour and efficacy of methodology proposed in the paper.

A comprehensive and sophisticated multi-algorithm coupled dynamic prediction model is proposed to address the intricate reality and stringent accuracy requirements of predicting tailings dam displacement. Firstly，by employing a time series decomposition model，the cumulative displacement is disaggregated into its trend and cyclical components. The trend term displacement is then forecasted using a Gaussian regression time series prediction model. Secondly，various Copula functions are employed to investigate the overall correlation between the inducing factors and the cyclical term displacement. Owing to the diverse influencing factors and strong nonlinearities associated with the cyclical term displacement，the MISSA-CNN-BiLSTM model is utilized for prediction. Lastly，the predicted trend term displacement from the Gaussian regression model and the predicted cyclical term displacement from the MISSA-CNN-BiLSTM model are merged. The results demonstrate a high degree of consistency between the predicted cumulative landslide displacements and the measured values，with a correlation coefficient of 0.996 and a root mean square error (RMSE) of 0.13 mm. The multi-algorithm coupled model，based on MISSA-CNN-BiLSTM，exhibits remarkable prediction accuracy and effectively captures step changes in tailings dam displacements.

In order to explore the damage characteristics of coal under different ultrasonic excitation time，the ultrasonic excitation test system，nuclear magnetic resonance imaging system，stereo microscope and automatic permeability test system of coal core were used to analyze the changes of pore structure，surface crack and permeability characteristics of coal under different ultrasonic excitation time. The results show that with the increase of ultrasonic excitation time，the change rates of T2 spectral peak area，total porosity，surface fracture area and permeability of small，medium and large pores in coal are linearly increased. When the coal body is excited by ultrasonic for 30 to 150 min，the change rate of T2 peak area of micro，medium and large pores increase from 14.60%，52.50% and 24.90% to 61.30%，145.50% and 235.70%，respectively. The total porosity change rate，surface crack area change rate and permeability change rate increased from 5.04 %，47.27 % and 41.67 % to 24.93 %，127.91 % and 208.33 %，respectively. Under the excitation of different ultrasonic time，the change rate of permeability of the coal and the change rate of total porosity and the change rate of surface cracking area are in line with the linear increasing relationship，with the increase of ultrasonic excitation time，the pore and crack modification effect of coal is enhanced，and the permeability can be improved.

In order to effectively reduce the risk of blind zones and lack of control in dust environment monitoring，optimize the node coverage control of the dust environment monitoring system in thermal power plants，and prolong the lifetime of WSN，an energy-saving optimization method based on improved genetic algorithms was proposed. Firstly，based on node coverage，total energy consumption of node deployment and total energy consumption of node communication and transmission，the network coverage quality objective function was constructed. Then，aiming at the problems of the local optimization and coding duplication existing in traditional genetic algorithms，the chromosome combination scheme of integer coding，the adaptive adjustment method of crossover and mutation probability and the elite retention strategy were proposed. Finally，the simulation comparison and analysis were performed to determine the optimized node number and distribution scheme. The results show that the improved genetic algorithm significantly improves the convergence speed. The number of iterations required is reduced to 20，and the fitness value is optimized by 52.18%. In the node deployment and coverage study，the optimized number of nodes is 42，the coverage rate is 97.28%，and the node dormancy rate is 76.19%，which effectively improves the energy-saving effect of the dust environmental monitoring system in the thermal power plant.

The S-FCN fire image detection method based on feature engineering was proposed to address the issues of high computational complexity and poor real-time performance of deep learning algorithms for fire image detection in complex backgrounds. Firstly，this method extracted color features from images in multiple color spaces and reduced the dimensionality of these features using mutual information. Secondly，the network structure of the deep learning model was simplified by using a single hidden layer of a fully connected network as its backbone. The color features in multiple color spaces can better represent fire smoke and flames，and reducing the dimensionality of color features in multiple color spaces effectively reduces the redundancy of input features. The single hidden layer fully connected network can significantly reduce the number of parameters during the model propagation process. Finally，this method was evaluated on a real and complex background fire image dataset. The experimental results show that the detection accuracy achieved by this method is 93.83%，and the real-time frame rate is 10 869 f/s. This method achieves high accuracy and high-speed fire image detection in complex scenes.

To achieve the accurate evaluation of steel structure fire resistance in the converter station，based on the possible standard，power and hydrocarbon fire scenarios，the thermal insulation mechanism of ultra-thin fireproof coating as well as its thermal insulation performance on steel members with different steel materials and different cross-section shape factors were investigated through micro-scale thermogravimetric-Fourier transform infrared (TG-FTIR) spectroscopy experiments and small-scale fire test furnace heat insulation experiments. The experimental results show that the faster the fire heating rate，the higher the peak mass loss rate and the higher the peak temperature of the fireproof coating. Different fire heating curves have no effect on the type of gases escaping from the thermal decomposition of fireproof coatings，but have an effect on the amount and peak temperature of escaping gases. Compared with the power and hydrocarbon fire，in the condition of standard fire，more gases are produced before 750 ℃，resulting in better expansion and heat insulation. Furthermore，the fire resistance of the ultra-thin fireproof coatings on the carbon steel and stainless steel at various cross-section shape factors at power and hydrocarbon fire is worse than that of standard fire，indicating poor heat insulation abilities.

To improve the city's resilience in responding to public safety risks，an integrated theoretical model of "spatial resilience" for urban public safety was proposed，and an urban public safety resilience evaluation index system was developed from the perspective of residents' perceptions. The entropy weight method was used to evaluate the public safety resilience level for Nanjing comprehensively. The results indicated that the overall score of Nanjing's public safety resilience was 6.851 5，and there was a structural imbalance in the resilience construction for the ternary space. Furthermore，social space safety resilience was the highest，followed by information space safety resilience，and physical space safety resilience was the lowest. The unbalanced development of urban public safety space resilience can be attributed to several factors including the lack of overall public safety resilience governance，insufficient investment in infrastructure and equipment maintenance，an imperfect social environmental risk monitoring system，and weak information security awareness and literacy among residents. Specific measures can be taken from the aspects of humanistic concepts，clear strategies，systematic thinking，and diversified cooperation.

In order to identify high-risk stations in urban rail transit systems and improve network resilience and operational safety，a performance function model was constructed to evaluate the network resilience by selecting network efficiency，average shortest path length and maximum connection sub-graph as indicators，and an evaluation method for critical stations was proposed，considering topological structure and passenger flow distribution equilibrium. Taking Nanjing Metro as an example，three cascading failure modes，descending critical degree，descending betweenness centrality and random sequence，are adopted. The characteristics of resilience degradation under different cascading failures are simulated respectively in the unweighted network and the weighted network. The results show that the critical stations are similar on weekdays and weekends. The resilience performance decreases most rapidly in the early stages of failures in descending order of critical stations. Compared with topology networks，the resilience index of passenger flow-weighted networks decreases faster in the early stages of cascading failures. Strengthening the control of critical stations when cascading failures do not spread widely can help reduce the loss of network resilience.