A statistical dependent failure detection method for n-component parallel systems. (November 2017)

• Record Type: Journal Article
• Authors: Ota, Shuhei; Kimura, Mitsuhiro
• Is Part Of: Reliability engineering & system safety. Volume 167(2017)
• Page Start: 376
• Page End: 382
• Publication Date: 2017-11
• Subjects: Cascading failure -- Parallel system -- Dependent failure detection -- K–S test; 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.2017.06.022
• Languages: English
• ISSNs: 0951-8320

Abstract

Highlights: A statistical detection method of dependent failure occurrence in parallel systems is proposed. The accuracy of the proposed method is demonstrated by simulation studies. A new stochastic model of dependent failure is presented by introducing the hazard rate switching. Abstract: Making a system redundant by combining identical components is a useful way to ensure a highly reliable system. However, the components of such systems may fail mutually, and if the components break down dependently, the reliability of the system decreases. Therefore, reliability analysis considering the dependence among the components is important in reliability assessment. This research proposes a statistical detection method of the dependent failure occurrence in n -component parallel systems that utilizes the failure occurrence times of the components. If we assume that the lifetime distribution of the components worsens if k -out-of- n components failed, the dependent failure occurrence can be found by identifying the change of the distribution. The performance of the proposed method is demonstrated by simulation studies.