A Physics of Failure Approach to Common Cause Failure Considering Component Degradation

• Published in: Proceedings. Annual Reliability and Maintainability Symposium pp. 1 - 6
• Main Authors: Smith, Reuel, Modarres, Mohammad
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.01.2020
• Subjects: common cause failure; Computational modeling; Degradation; Failure analysis; life-stress model; Mathematical model; physics-of-failure; Reliability; reliability simulation; System reliability; Temperature distribution
• ISSN: 2577-0993
• DOI: 10.1109/RAMS48030.2020.9153651

This paper outlines the development of a new approach to reliability analysis of aging systems involving common cause failure of redundant components, that are subject to degradation failure mechanisms described by physics-of-failure (PoF) models; that is, the approach considers the presence of various failure mechanisms that increase the hazard rate of the system's redundant components. The approach produces a more accurate time-to-failure and system reliability, including better applications to prognosis and health management. A MATLAB simulation was designed to model the PoF of a particular multi-component system, where the underlying degradation failure mechanisms served as the coupling common cause failure mechanisms (CCCFM). The simulator developed to test this Common Cause Physics of Failure (CCPoF) model was demonstrated using a two-component system. The simulator required identification of failure mechanisms, where the temperature-based failures CCCFM Arrhenius life-stress model served as an example. The simulator randomly generated realizations of CCCFM component failure times and the time of failure of the system itself. Finally, the traditional time-based beta factor common cause failure (CCF) model and system failure probability were obtained for comparison to the CCPoF approach. This simulation procedure was repeated for several levels of ambient temperature to determine any relevant trends and findings.