Reliability assessment of film capacitors oriented by dependent and nonlinear degradation considering three-source uncertainties

Abstract

Reliability-oriented problems of film capacitors are very crucial to ensure the reliable and healthy operation of power electronic converters. Capacitance and equivalent series resistance (ESR) are the typical performance characteristics (PCs) of capacitors, and their degradation processes also show certain dependence, which was almost ignored in previous researches. Besides, the degradation paths of these PCs are often nonlinear and affected by three-source uncertainties including temporal uncertainty, individual heterogeneity and measurement errors, which are rarely simultaneously considered in the existing dependent degradation models. Regardless of ignoring the dependence, nonlinearity or uncertainties in the degradation process, the reliability assessment result will be inaccurate. In this article, a new reliability assessment method for film capacitors is put forward oriented by dependent and nonlinear degradation considering three-source uncertainties. First, a random-effect nonlinear Wiener-based model with measurement errors is developed to explicitly characterize the degradation processes of capacitance and ESR. Subsequently, the appropriate Copula function is selected and employed to capture their dependence. Finally, an accelerated degradation test (ADT) on film capacitors combined with comparative analysis fully highlights the superiority of the new reliability assessment method.