Belief Reliability Modeling for a System with Degradation Interaction Effect

Abstract

Complex systems often have multiple performance characteristics deteriorating with time. Due to the influence of some common factors (such as internal structure, application experience, environmental operating conditions, etc.), the dependence among multiple performance characteristics should be considered. Aiming at the problem of how to describe the coupling relationship of multiple degradation processes, a reliability modeling method for a system with bivariate dependent degradation processes based on the belief reliability theory and degradation interaction effect is proposed in this paper. Firstly, the degradation interaction effect is employed to describe the dependence between two performance characteristics during their degradation evolution process. Secondly, the degradation equation, margin equation and matric equation of two performance characteristics are established according to the belief reliability theory framework. Then, the statistical inference method based on genetic algorithm is used to estimate the unknown parameters. Finally, a real application on the tuner is provided to illustrate the performance of the proposed model and statistical estimation method.