A novel interpretable model of bathtub hazard rate based on system hierarchy. (December 2022)

• Record Type: Journal Article
• Authors: Du, Yi-Mu; Sun, C.P.
• Is Part Of: Reliability engineering & system safety. Volume 228(2022)
• Publication Date: 2022-12
• Subjects: Reliability theory -- Hazard rate function -- Complex system -- Bathtub shape; 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.2022.108756
• Languages: English
• ISSNs: 0951-8320

Abstract

Abstract: The degradation of a functioning system can be characterized using the hazard rate functions (HRFs) in the context of reliability theory. The bathtub hazard rate (BHR) is a widely-seen form of HRF for many practical systems when considering the system as a whole. Many previous studies attempt to propose different predictable models at system level for BHRs. However, since these models are usually not interpretable, they can only deal with the system-level lifetime data. If the data could not bring explicit HRF shape due to the limitation of its amount, then it is hard to select a proper model. To offset the shortcoming of system-level models, we study the underlying mechanism in a multiple-component scenario and provide a model based on hierarchy–structure systems to infer the appearance of BHR via their structure induced failure modes. The novel model is interpretable and renders a steady BHR at system level which could be a theoretical support for the BHR in reliability engineering. The proposed model provides several applications in different practical conditions, such as the prediction of the HRF's shape with only information of structure, inference of structure with lifetime data and fusing information from different levels. Highlights: The study provides an extensible yet interpretable model to construct steady BHR for complex systems. This study proposes a novel mechanics-based method to infer the appearance of BHR for complex systems via their structures. This study offers a new systematically approach to analyze the hazard rate function for the system of systems.