关键词: 动态贝叶斯网络, 故障树分析法, 气化炉, 风险分析

Abstract: Aiming at the shortcomings of dynamic risk analysis in large gasification units, a method of dynamic Bayesian network based risk analysis of gasifier feed system was proposed. The fault tree model was established by using the relevant data of failure forms of each unit of gasifier feeding system, and the fault tree model was transformed into a Bayesian network model. The bayesian network model was optimized by K⁃2 algorithm, and the influence of common cause failure on the system was processed by β factor method. Considering the influence of maintenance factors on the failure rate of the system at each moment, the accuracy and feasibility of the proposed method were verified by using the Bayesian network analysis software GeNIe and Monte Carlo. The results show that the dynamic Bayesian network has the advantages of accurate dynamic risk analysis and the ability to identify the weak links in the gasifier feed system risk analysis. According to the analysis results, a reasonable reliability allocation strategy can be obtained.

Key words: Dynamic bayesian network, Fault tree analysis, Gasifier, Risk analysis