Degradation and reliability of multi-function systems using the hazard rate matrix and Markovian approximation. (February 2022)

• Record Type: Journal Article
• Authors: Zhou, Daoqing; Sun, C.P.; Du, Yi-Mu; Guan, Xuefei
• Is Part Of: Reliability engineering & system safety. Volume 218:Part B(2022)
• Publication Date: 2022-02
• Subjects: Reliability -- Hazard rate matrix -- Markovian approximation -- Multi-function systems -- Degradation process; 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.2021.108166
• Languages: English
• ISSNs: 0951-8320

Abstract

Abstract: The study presents a new method to model the degradation and reliability of multi-function complex systems using the hazard rate matrix and Markovian approximation. The system is hierarchically decomposed into a set of states according to the states of its functions, and the hazard rate matrix is proposed to describe the failure rates of the functions. By using Markovian approximation, the elements of the hazard rate matrix can be expressed as a set of dynamical equations involving the failure information about the functions. Consequently, the dynamical behavior of the degradation for the whole system and its functions can be determined using the observed failure data of those functions instead of the lifetime data of the whole system, eliminating the need for lifetime testing data of the whole system in statistical-based reliability assessment. The overall method is illustrated using an example problem, and is further applied to a power system degradation problem to demonstrate its usage in realistic engineering applications. Highlights: Hazard rate matrix is proposed to model the degradation of complex systems. Hazard rate matrix and system state transition rate matrix are correlated. Method yields close-form solutions under Markovian condition. System degradation can be resolved using function failure times.