A generalized system reliability model based on survival signature and multiple competing failure processes

Abstract

Degradation-based system reliability analysis has been extensively conducted, but the components in a system are assumed to experience similar degradation and shock processes, neglecting actual failure mechanisms. However, multiple types of components in a system may work under different operational conditions and break down due to different failure mechanisms. Hence, a new generalized reliability model is proposed for systems with arbitrary structures experiencing multiple degradation and shock processes, including pure degradation processes (DPs), independent and dependent competing failure processes (CFPs). In this work, the Tweedie exponential-dispersion (TED) process is utilized to describe multiple degradation processes of the components, which contains the Wiener, Gamma, inverse Gaussian, and other processes as special cases. Based on multiple DPs and CFPs, a generalized reliability model is established by utilizing the structure analysis method, the survival signature, which allows the proposed method to be applied to various structural systems. Finally, an example of an automotive braking system with four types of components experiencing multiple DPs and CFPs is applied to illustrate the proposed model.