Reliability assessment of mechatronic systems considering multi covariates with system topology. Issue 1 (15th September 2021)

• Record Type: Journal Article
• Authors: Lin, Shuai; Jia, Limin; Zhang, Hengrun; Wang, Yanhui
• Is Part Of: Systems engineering. Volume 25:Issue 1(2022)
• Page Start: 68
• Page End: 90
• Publication Date: 2021-09-15
• Subjects: covariates -- hazard function -- mechatronic system -- system reliability -- system topology; Systems engineering -- Periodicals; 620.0011
• DOI: 10.1002/sys.21602
• Languages: English
• ISSNs: 1098-1241

Abstract

Abstract: Reliability assessments of mechatronic systems, considering covariates, are becoming more complex not just as a result of the integration of more diverse technologies into such systems but also because of increasing complexities in system topology; but traditional reliability assessment models with covariates tend to ignore the influence of system topology. This paper addresses such an exclusion and proposes a novel methodology dedicated to the analysis and quantification of system reliability using system topology and multiple covariates. There are essentially two new aspects contributing to the overall methodology. Firstly, using network theory, the system topology is abstracted as a network; in which network nodes represent a mechatronic system's minimum maintenance units, and the connections between them are regarded as edges. Secondly, the resultant topological network is combined with the inherent properties of the minimum maintenance units to construct the system baseline hazard function and considering the associated multi covariates from the perspective of failure propagation to provide a foundation for assessing system reliability. To validate the proposed methodology, by way of a case study, it was applied to the reliability assessment of a traction power supply system for a high‐speed train. The results, compared with three other assessment approaches, suggest that the proposed methodology provides for significantly more accurate assessments while requiring less computational resources. Furthermore, it can be flexibly extended to many other mechatronic systems.