Failure behavior modeling and reliability estimation of product based on vine-copula and accelerated degradation data

Aiming at the characteristics of failure mechanism coupling of high-reliability and long-lifetime complex products, a method of multivariate failure behavior modeling and reliability assessment is proposed based on vine-copula and accelerated degradation data. The method, considering the coupling modeling under accelerated stress and the relevant acceleration mechanism consistency test, can not only clearly describe the multivariate coupling relationship of the product, but also can carry out the reliability assessment of the product within the affordable time and cost. The failure behavior modeling method consists of two parts which are degradation behavior and correlation relationship among variables. Firstly, degradation behavior description of failure behavior modeling is obtained from, the degradation data of product performance parameters based on the accelerated degradation test, the degradation path of each performance parameter by utilizing drift Brownian motion, then the implementation of accelerated consistency test ensures the true reliability information. Furthermore, the correlation description of failure behavior modeling is constructed by Vine-Copula based multivariate damage coupling modeling method describing the pairwise correlation relationship. Afterward, the exact solution of reliability joint distribution of the product with non-strictly monotonic degradation is figured out by Vine-Copula and the equation of reliability estimation based on conditional probability. Finally, the feasibility of the method is verified by utilizing smart electricity meter as the numerical example.