A Copula-based Degradation Modeling and Reliability Assessment

In this paper, a novel system reliability assessment model with two dependent performance characteristics (PCs) is proposed via the marginal distribution functions and copula theory. Firstly, the best fitted marginal distribution function is obtained by using the approximation method and the Anderson Darling test statistic. Secondly, the dependent of the two PCs is described by different copula functions, and the unknown parameters of the proposed model are obtained by using an inference for margins method. Finally, a numerical example about an actual train wheels wear degradation data is given to demonstrate the usefulness and validity of the proposed model and method. Numerical results show that ignoring the dependence between two PCs may result in different reliability conclusion.