Degradation model analysis of two-component series and parallel systems

Abstract

ABSTRACTMany high-reliability and long-life products to form a system in series or parallel. Based on the assumption that the degradation process of the performance characteristics is Wiener procession, this article proposes a degradation model of the series and parallel systems. Here, a bivariate constant-stress degradation model is established, assuming Wiener procession and utilizing the copula function to describe the correlation among the performance characteristics. The adoption of an inference function for margins method for estimating the model's parameter ensures the accuracy of the estimation results as well as reduces the difficulties of parameter estimation effectively. In the numerical example research, we estimate the parameters of the degradation models of Gaussian Copula and Frank Copula using the rubidium spectral lamp as an example. Furthermore, we also adopt the Monte Carlo simulation method on the effects of copula function's estimation. We find that the Frank Copula more applicable than Gaussian Copula. For the two-component series and parallel systems composed by the rubidium spectral lamp, we get the reliability curve, the failure probability density curve and the MTBF value of the series system and parallel system. These curves are compared and analysed to achieve the conclusions.