Novel Cumulative Degradation Approach to Predict Components Failure Rates

Abstract

This paper describes a novel approach for predicting failure rates for components and hardware systems. A physics-of-failure-based methodology is provided to predict the degradation rate of a population using a Monte Carlo approach. The methodology relies on an empirical equation describing the degradation of a performance metric, a degradation consistency condition and a technique to account for cumulative degradation across multiple temperatures. Two case studies are given to illustrate the methodology.