Binary decision diagram-based methods for risk assessment of systems subject to propagated failures and parametric uncertainties

Propagated failures between components mean that the failure of one component may cause the failures of other components within the system. Because propagated failures contribute greatly to the system unavailability, it is important to take into account propagated failures between components when modeling systems. Sometimes, it may be difficult to obtain the precise values of some parameters due to many reasons, such as the insufficiency of historical data and the randomness of statistical data. In this case, interval values can be given instead of precise values to represent the parametric uncertainty related to the values of parameters. This paper proposes explicit and implicit binary decision diagram-based methods to model systems subject to propagated failures and to evaluate the failure probability of systems under parametric uncertainty. The proposed methods are applied to a near head-to-head railway accident to evaluate the occurrence probability of accidents.