A fuzzy-based reliability approach to evaluate basic events of fault tree analysis for nuclear power plant probabilistic safety assessment

ABSTACK
Fault tree analysis has been widely utilized as a tool for nuclear power plant probabilistic safety assess-
ment. This analysis can be completed only if all basic events of the system fault tree have their quanti-
tative failure rates or failure probabilities. However, it is difficult to obtain those failure data due to
insufficient data, environment changing or new components. This study proposes a fuzzy-based reliabil-
ity approach to evaluate basic events of system fault trees whose failure precise probability distributions
of their lifetime to failures are not available. It applies the concept of failure possibilities to qualitatively
evaluate basic events and the concept of fuzzy sets to quantitatively represent the corresponding failure
possibilities. To demonstrate the feasibility and the effectiveness of the proposed approach, the actual
basic event failure probabilities collected from the operational experiences of the David–Besse design
of the Babcock and Wilcox reactor protection system fault tree are used to benchmark the failure prob-
abilities generated by the proposed approach. The results confirm that the proposed fuzzy-based reliabil-
ity approach arises as a suitable alternative for the conventional probabilistic reliability approach when
basic events do not have the corresponding quantitative historical failure data for determining their reli-
ability characteristics. Hence, it overcomes the limitation of the conventional fault tree analysis for
nuclear power plant probabilistic safety assessment.