Risk analysis of hydrogen refueling station leakage and explosion accidents using dynamic Bayesian networks
We propose a novel and dynamic risk analysis method for addressing the uncertainty,complexity,and time series dynamics associated with hydrogen leakage and explosion accidents in hydrogen refueling stations.This method aims to provide a more thorough and effective approach to assessing the risk factors involved in such accidents.This approach is grounded in dynamic Bayesian networks and employs a fusion technique that combines GM(1,1)grey prediction and fuzzy set theory to build a prior probability model.When there is a lack of risk factor probability data,it can forecast the prior probability of risk factors by analyzing the temporal variation pattern of the hydrogen refueling station's risk avoidance capability.Introducing the Leaky Noisy or gate model and Markov hypothesis can aid in addressing the impact of omitted risk factors and handling the uncertainty of conditional probabilities in dynamic Bayesian networks.This approach can also make the model more objective and suitable for short-term prediction of accident evolution under actual working conditions.The Dynamic Bayesian Network(DBN)model was developed to analyze the risk of leakage and explosion accidents at a specific hydrogen refueling station.The accuracy and reliability of this method were confirmed through the application of the three axioms theorem and risk factor sensitivity analysis.Through quantitative analysis,diagnostic analysis,sensitivity analysis,and influence strength analysis,we have determined the likelihood of accident consequences like"jet fire"and its time-sequence change curve.Additionally,we have assessed the probabilities of catastrophic accidents resulting from leaks in various work units and at different hydrogen leakage flow rates.Key risk factors,such as"pipeline corrosion prevention failure"and highly sensitive risk factors like"high-pressure hydrogen pipeline failure",have been identified.Moreover,we have established the key causal chains of leakage and explosion accidents.Utilizing the risk factor impact model for hydrogen refueling stations,tailored recommendations are put forth for enhancing equipment health management,implementing pipeline anti-corrosion measures,and enhancing other risk prevention and control strategies.
safety engineeringhydrogen refueling stationsleakage and explosion accidentsrisk analysisDynamic Bayesian Network(DBN)
万方数据
维普
