Two decades of statistical approach in reliability prediction of power converters

Abstract

Various statistical methods are employed to predict the reliability of electronic components and monitor improvements in reliability. This article conducts a comprehensive study of statistical approaches for predicting the reliability of power converters, as reported in the literature. The analysis focuses on three statistical methods: MIL-HDBK-217F, FIDES, and RIAC. Among these, the military handbook MIL-HDBK-217F has been a dominant tool for failure analysis over the past two decades. The general formulas used for failure rate analysis of power converter parts with each of the three approaches are presented. The study also reviews and discusses failure rate analysis for components and modules in power converters, both with and without fault-tolerant capabilities, as documented in the literature using MIL-HDBK-217F, FIDES, and other statistical approaches. Finally, the paper outlines the conclusions drawn from the study and highlights the main limitations of the three methods. This article serves as a resource for researchers working on fault analysis in power converters.