A combined physics of failure and Bayesian network reliability analysis method for complex electronic systems

Complex electronic systems have a structures that can lead to coupling failure mechanisms and difficulties in collecting measured data. These issues increase the difficulty of reliability analysis. Current reliability research methods cannot effectively solve the above problems. In this paper, we propose a new approach that combines a physics of failure (PoF) method and a copula Bayesian network to assess complex electronic systems. The proposed approach improves the defects of PoF methods and traditional Bayesian networks when applied to the reliability analysis of complex electronic systems. A copula Bayesian network is used to realize the dependent failure modeling of modules or components for interlevel failure and intra-level failure. The PoF method addresses the difficulty in obtaining measured data. This proposed approach is applied to the reliability analysis of the integrated processor system in communication equipment. The key impacted subsystems and devices are analyzed from three aspects-qualitative analysis, forward inference and backward inference-and the corresponding failure life distributions are calculated. This method can guide the improvement of system reliability and system maintenance. (C) 2021 Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

Automated summary

This paper introduces a new approach that combines the PoF method and copula Bayesian network to improve the reliability analysis of complex electronic systems. The method entails qualitative analysis, forward inference, and backward inference to analyze key subsystems and devices, and calculate failure life distributions.