Research On Dynamic Risk Analysis And Accident Consequence Control For Third-Party Damage Of Natural Gas Pipelines

In the context of the “National One Network” strategy,the total length of existing natural gas pipelines in China has exceeded 110,000 kilometers,which also means that the pipeline will face a more complex surrounding environment and more external influences.Currently,the proportion of pipeline failures caused by third-party activities has exceeded the proportion of failures caused by corrosion accidents.Due to the strong randomness of third-party destructive behavior,it has seriously affected the normal operation of pipelines.Therefore,this study evaluates the disaster-bearing capacity of case pipelines,dynamically deduces the probability of accident risk factors and potential accident consequences for pipeline sections with poor disaster-bearing capacity and proposes corresponding control measures based on risk acceptance standards.The main research contents and conclusions are as follows.(1)According to natural gas pipeline accident reports,relevant materials,etc.,a total of45 factors influencing third-party damage were identified and analyzed.The accident tree model(FTA)and event sequence diagram(ESD)are used to describe the development process of risk accident.Therefore,the development of a complete pipeline failure scenario of risk factors-failure accidents-potential consequences is realized.(2)A third-party damage-bearing capacity assessment index system is established and an example analysis is carried out.After a series of preprocessing operations on the index data,the comprehensive disaster-bearing capacity level of the pipeline was obtained by cluster analysis by Gaussian mixed clustering(GMM).Particle Swarm Optimization(PSO)is used to improve the parameter setting defects of traditional PNNs.The PSO-PNN model is established to evaluate the disaster-bearing capacity level of each pipe section,and the accuracy of the final evaluation result is 98%.(3)Based on Bayesian network(BN),a third-party destruction dynamic risk analysis model is established.The prior probability of basic events is obtained through database information and fuzzy comprehensive evaluation,and 6 key hazard events are derived by probability update and probability adaptation.Based on the hidden Markov model theory(HMM),the BN model is extended to a dynamic Bayesian network model(DBN)to complete the probability prediction of risk accidents.(4)For the three consequences of structural damage,combustion explosion and explosion hazard,the acceptable range of personal and social risks is formulated with reference to the risk acceptance standards.The results show that only structural damage events show that their risk is acceptable when the disaster bearing level is 5,while the risk of the other two potential accident consequences exceeds the acceptable standard.So,certain control measures need to be taken to reduce the risk.The results show that the PSO-PNN model constructed in this paper can accurately evaluate the disaster-bearing capacity of pipelines,and the dynamic risk deduction results based on the DBN model are basically consistent with the actual situation.The model constructed by this institute has certain guiding value for the risk prevention and accident control of third-party damage in natural gas pipelines.