Failure mechanism dependence and reliability evaluation of non-repairable system. (June 2015)

• Record Type: Journal Article
• Authors: Chen, Ying; Yang, Liu; Ye, Cui; Kang, Rui
• Is Part Of: Reliability engineering & system safety. Volume 138(2015:Jun.)
• Page Start: 273
• Page End: 283
• Publication Date: 2015-06
• Subjects: Failure mechanism dependence -- Trigger -- Acceleration -- Reliability evaluation -- Non-repairable system; Reliability (Engineering) -- Periodicals; System safety -- Periodicals; Industrial safety -- Periodicals; Fiabilité -- Périodiques; Sécurité des systèmes -- Périodiques; Sécurité du travail -- Périodiques; 620.00452
• DOI: 10.1016/j.ress.2015.02.002
• Languages: English
• ISSNs: 0951-8320

Abstract

Abstract: Reliability study of electronic system with the physics-of-failure method has been promoted due to the increase knowledge of electronic failure mechanisms. System failure initiates from independent failure mechanisms, have effect on or affect by other failure mechanisms and finally result in system failure. Failure mechanisms in a non-repairable system have many kinds of correlation. One failure mechanism developing to a certain degree will trigger, accelerate or inhibit another or many other failure mechanisms, some kind of failure mechanisms may have the same effect on the failure site, component or system. The destructive effect will be accumulated and result in early failure. This paper presents a reliability evaluation method considering correlativity among failure mechanisms, which includes trigger, acceleration, inhibition, accumulation, and competition. Based on fundamental rule of physics of failure, decoupling methods of these correlations are discussed. With a case, reliability of electronic system is evaluated considering failure mechanism dependence. Highlights: Five types of failure mechanism correlations are described. Decoupling methods of these correlations are discussed. A reliability evaluation method considering mechanism dependence is proposed. Results are quite different to results under failure independence assumption.