Reliability Analysis of Probabilistic Competing Failure Systems with Both Selective and Global Failure Propagation Effects

This paper investigates the reliability modeling of systems subject to probabilistic competing failure behaviors with complicated failure propagation effects. To be specific, besides local failure (LF) that only affects the component itself, a component can also experience propagated failure with global effect (PFGE) leading to the system-wide failure, and propagated failure with selective effect (PFSE) that affects the set of other components. There exists a probabilistic functional dependence dynamic between system components, where some components (referred to as the probabilistic-dependent components) functionally depend on other components (designated as trigger components) and can be isolated by the trigger component failure in a probabilistic manner. Trigger component LF and PF of probabilistic-dependent components compete in the time domain. Different occurrence orderings of these component failures can lead to dramatically different system states. However, the existing reliability assessment methods are not applicable to addressing such probabilistic competing failure behaviors with considering both PFGEs and PFSEs in the system reliability analysis. A novel combinatorial reliability methodology is presented to tackle this issue with its applicability and effectiveness being demonstrated through step-by-step reliability analysis on a smart home sensor system. The proposed method is verified and the methodology complexity is discussed by comparing with the Markov method.