Reliability Analysis and Maintenance Policy for Complex Systems

Abstract

With the advances in emerging technologies, systems are becoming increasingly complex, which brings enormous challenges for reliability engineering. On the other hand, due to lack of information for complex systems, how to appropriately implement maintenance policies remains an open issue. The thesis, comprising five studies, investigate the reliability analysis and maintenance policy for complex. Our work can be classified into three categories: condition-based maintenance (CBM) policy, maintenance scheduling for multi-component systems and applications on specific systems. Our research contributes to safety guarantee and operating management for emerging technologies and complex systems.
Condition-based maintenance policy. In the area of CBM policy, two studies are conducted to enrich the research of CBM. We construct a maintenance model considering side effect of aging and degradation, which is characterized as the age- and state-dependent operating cost in maintenance framework. The age- and state-dependent operating cost occurs when the system is operating in a deteriorated state, even if the system is still functioning. The system is subject to a continuous-time degradation process. The Wiener process is used to characterize the evolution of the system states. By taking advantage of the independent increment of the Wiener process, we model the CBM policy as a Markov decision process. The system is subject to periodic inspections. If the system has not failed at inspection, decision has to be selected among three maintenance actions: preventive replacement, repair and wait until the next inspection. The system is replaced when its degradation level exceeds a pre-specified threshold at inspection. The structural property of the optimal maintenance decision is investigated in depth and a monotone control limit policy is shown as the optimal strategy.
Besides, we propose a degradation-integrated failure model, which argues that degradation itself does not directly lead to system failure, but increase the failure rate. CBM models are established for systems subject to aging and cumulative damage. PHM is used to model the joint effect of aging and cumulative in the framework of failure rate. The effect of cumulative damage is modeled as the stochastic covariate in the PHM framework. The system is subject to periodic inspection, which is assumed to be perfect. At inspection, maintenance actions are carried out based on the observed condition information. Optimal maintenance policies are obtained by minimizing the long-run cost rate. Specifically, two CBM models are developed by assuming respectively known distribution parameters and unknown distribution parameters. In the case where the distribution parameters are unknown, the parameters have to be estimated and updated at each inspection, and maintenance decisions are made subsequently.
Maintenance scheduling for multi-component systems. We develop a maintenance policy for a multi-component system subject to hidden failures. The objective of this study is to determine the inspection intervals for each component such that the long-run cost rate is minimized. Here, we focus on the effect of economic dependence since economic dependence is the most common interaction in a multi-component system. With respect to the maintenance optimization of the multi-component system, the difficulty lies in the computational complexity. Actually, optimization of multi-component maintenance scheduling with economic dependence has been proven to be an NP-complete problem. To deal with the intractability of this problem, we adopt a heuristic method named “base interval approach”, which reschedules the inspection interval of each component as an integer multiplier of a base interval, so as to share the common downtime cost and reduce maintenance cost. In addition, we investigate the performance of the base interval approach, where the effectiveness of proposed approach is demonstrated both analytically and numerically.
Applications on specific systems. With respect to applications of the maintenance policies, we establish reliability and maintenance models for two specific systems: load-sharing systems and mission-oriented systems. We construct reliability models for load-sharing systems subject to degrading components so as to arrive at a preventive maintenance strategy. At first we obtain some preliminary insights by modeling system reliability with constant load. Next we build a reliability model assuming varying loads. Specifically we consider the scenario where the load is random and has a cumulative impact on the system. Finally, we utilize the proposed reliability models to arrive at preventive maintenance decisions.
In addition, we develop an imperfect maintenance model for mission-oriented systems subject to degradation and external shocks. Instead of availability constraint, reliability constraint is adopted as a requirement during mission operation. We select reliability over availability as the constraint because reliability is more appealing than availability for a safety-critical system where failure of a mission leads to huge losses. In particular, we consider multiple dependent competing failure processes where either external shocks or natural degradation can lead to system failure. The degradation process is modeled as a Wiener process and the shocks are assumed to arrive according to Poisson process. The optimal maintenance policy is achieved by minimizing the long-run cost rate.

Date of Award	1 Aug 2017
Original language	English
Awarding Institution	City University of Hong Kong
Supervisor	Way KUO (Supervisor) & Min XIE (Co-supervisor)