Selective maintenance on a multi‐state system considering maintenance task assignment and operating cost. Issue 2 (23rd November 2021)

• Record Type: Journal Article
• Authors: Sun, Yao; Sun, Zhili; Zhou, Jie
• Is Part Of: Systems engineering. Volume 25:Issue 2(2022)
• Page Start: 157
• Page End: 172
• Publication Date: 2021-11-23
• Subjects: maintenance task assignment -- multi‐state system -- operating cost -- selective maintenance; Systems engineering -- Periodicals; 620.0011
• DOI: 10.1002/sys.21608
• Languages: English
• ISSNs: 1098-1241

Abstract

Abstract: For some engineering systems where there is a finite break between two consecutive missions, maintenance activities can only be performed during the break, and the budget is also limited. Due to limited resources, it is necessary to select a subset of the maintenance activities to be performed, with the aim of meeting the next mission requirement. This is termed as selective maintenance. We propose a selective maintenance model for a multi‐state system in which both maintenance task assignment and operating cost are considered. The algorithm of maintenance task assignment is proposed inspired by the 0–1 knapsack problem to help fulfill the maintenance optimization. The correctness and computationally effectiveness of this algorithm are validated. With the degradation of each component modeled as a homogenous Markov process, the operating cost of the system is modeled based on the Markov reward process and is treated as part of the total cost. Additionally, the randomness of the operating time is considered and the necessity of this consideration is illustrated. The proposed model is applied to a coal transportation system to demonstrate its effectiveness, and the effects of ignoring maintenance task assignment and operating cost are analyzed. The results show that it is essential to take the maintenance task assignment into consideration, especially when the time constraint plays a major limiting role, as failure to do so may lead to poor system optimization. Furthermore, not considering the operating cost will underestimate the requirement of the resources, more so when a higher reliability is required.