基于模糊动态贝叶斯网络的加氢站氢气泄漏动态风险评估方法OA

This paper proposes a risk assessment method for hydrogen leakage at hydrogen refueling stations based on fuzzy dynamic Bayesian networks.First,the main risk factors causing hydrogen leakage at refueling stations are analyzed,and a corresponding bowtie model is constructed.Based on this,the bow tie model is transformed into a dynamic Bayesian network structure using a mapping algorithm.To reduce bias caused by subjective judgment,this study uses a combination of 1expert scoring and fuzzy set theory to determine the prior probabilities of basic events.Simultaneously,considering the dy-namic evolution of risk factors,the concept of time slices is introduced into the model,and conditional probabilities are calculated using a leaky noisy OR gate model.Finally,an ap-plication analysis is conducted using a hydrogen refueling station in an industrial park as a case study,and the results show that the method has good practicality.Calculations based on the fuzzy dynamic Bayesian network show that the initial probability of hydrogen leakage at this refueling station is 7.896×10-3,while by the tenth month,the leakage probability has increased to 9.48×10-3.