Probability and Frequency Derivation Using Dynamic Fault Trees

Abstract

Dynamic Fault Tree (DFT) is suitable to capture functional and dynamic dependencies among events leading to system failure. There exist several approaches for analyzing DFTs, each of them suffers from drawbacks that limit its practical use. This paper presents a comprehensive methodology based on the structure function determination. It consists in performing both qualitative and quantitative analysis to establish failure sequences and to calculate the likelihood (probability and frequency) of dynamic systems failure. Failure sequences are extracted from the minimal canonical form of the top event and allow getting disjunctive failure sequences. From the quantitative point of view, the probability and frequency of any DFT top event are developed to make the proposed model able to quantify the failure probability and frequency of dynamic systems and the frequency of accident scenarios. The proposed procedure is applied on two benchmark examples: the first one is a safety‐related system while the second one is an accident scenario with dependent protection layers. In order to check the validity of the proposed method, the derived numerical results are compared with those obtained from Markov Chains models. © 2018 American Institute of Chemical Engineers Process Saf Prog 37: 535–552, 2018