Review of Application on Dynamic Fault Tree Method in Nuclear Power Plants

Abstract Fault tree analysis (FTA) is one of the most important methods of probabilistic risk assessment (PRA). The fault state of the system is taken. While traditional FTA is based on static failure model. FTA is not applicable for systems that include redundant, sequence-related systems. At the same time, nuclear power plants (NPPs) contains a large number of redundant equipment, and FTA is difficult to solve these dynamic problems. Therefore, it is necessary to use dynamic fault tree analysis (DFTA) for PRA. In DFTA research, the modular analysis method was first proposed. The modular method divides the dynamic fault tree into a dynamic fault tree and a static fault tree. Among them, the dynamic fault tree is analyzed using a Markov chain, and the static fault tree is studied using a binary decision diagrams method. However, the shortcomings are that when the system is complicated, the information explosion in the Markov chain is appeared. To solve this problem, a dynamic fault tree is transformed into a Bayesian network. At the same time, to verify the feasibility of the method, Monte Carlo random sampling was used to evaluate the method. Other methods are relatively infrequently studied. In this paper, firstly, status of dynamic fault trees has been investigated. Secondly, each method is analyzed and the problems of dynamic fault tree are described. Finally, a survey and analysis on the dynamic fault tree is finished, and the main problems of the dynamic fault tree are: information explosion; the lack of commercial software to apply to engineering. Through this review, we hope to play a certain guiding role in the subsequent research on dynamic fault trees.