A new Bayesian network approach to solve dynamic fault trees

We explore the usage of Bayesian networks (BNs) for reliability modeling and analysis of dynamic systems. Dynamic system components exhibit complex behaviors and interactions, making combinatorial models inappropriate to solve them. Markov chains, and their extensions, have been widely used to model such systems. However, the infamous state space explosion problem greatly limits their application. We propose a novel reliability modeling and analysis framework based on temporal BNs. Our main concern in this paper is to define a BN-based modeling framework to solve dynamic fault trees and deal with the state space explosion problem encountered in Markov chain-based solutions. In summary, we demonstrate that a BN-based reliability framework, as defined in this work, is a powerful potential solution to address some of the problems existing in current reliability modeling and analysis formalisms. In particular, we show how the framework is capable of modeling and analyzing various kinds of system components' behaviors and dependencies and more importantly how the BN approach can be used as an alternative solution of a dynamic fault tree model without resorting to the Markov chain generation and solution.