Risk analysis of boiler overpressure explosion based on complex network and fuzzy Bayesian inference

To improve the risk of overpressure explosion in industrial boilers and solve the lack of uncertainty treatment in risk analysis, fuzzy Bayesian network (FBN) and complex network (CN) theory are applied to carry out risk analysis of overpressure explosion in industrial boilers. First of all, based on the statistical analysis of 586 industrial boiler accident cases, 37 typical accident causative factors are derived based on the Hazard and Operability Analysis (HAZOP), and the disaster chain is extracted from the "man-machine-environment-management" view. And then the nodes and edges of the disaster chain are constructed by Bayesian network (BN) topology, and the triangular fuzzy set theory is used to determine the a priori and a posteriori probability of accidents, which are used to reason out the failure probability of boiler overpressure explosion accident causally and deduce the accident chain evolution. Finally, by the use of CN node theory, the key nodes for cutting the chain of the industrial boiler overpressure explosion accident are comprehensively determined. The results show that using a FBN can accurately quantify the failure probability of boiler overpressure explosion accidents and the accident chain evolution path, while utilizing the complex network can identify the critical nodes of the broken chain. This risk analysis method can provide a decision-making basis for the safety management of industrial boilers.