چكيده به لاتين
The maintenance modeling aims to determine an optimal maintenance policy for a system. Various factors may affect an optimal maintenance policy for a multi-component system such as the type of dependencies between components, the failure modes of components, the inspection type, the kind of maintenance actions, the length of time horizon, the sort of optimization criteria, etc. First, the definitions, basic concepts, and common maintenance policies are briefly outlined. Then, inspection models for detecting hidden failures and potential defects are investigated according to the failure modes of components.
It is notable that most of the developed maintenance models focus on a system subject to either the hidden failures or two-stage revealed failures. The lack of simultaneous consideration of both hidden and two-stage revealed failures with stochastic dependence in the literature on inspection modeling made us conduct this study. Some components may develop a hidden failure, an occurrence of which neither stops the system nor accelerates the other components’ deterioration. However, the hidden failure reduces the system’s performance by increasing its operating costs. On the other hand, the other components may experience three states: healthy, defective and revealed failure. Any revealed failure of each component not only stops the system but also generates a shock to all the other ones, which increases their occurrence rate of hidden, defect and revealed failures.
First, a block-based maintenance policy based on periodic and opportunistic inspection is mathematically formulated to take advantage of economic dependence. Under this policy, in addition to the periodic inspections, the system is also inspected at revealed failures moments of each component to detect and fix both defects and hidden failures on all the other components. Minimal repair can fix the defective components; however, the failed components should be immediately replaced. Then, the proposed policy is developed by taking into account two types of maintenance actions (minimal repair and replacement) and two inspection errors for each component of the system. The goal is to find the optimal periodic inspection interval such that the expected total cost is minimized over a finite time horizon.
In the proposed approach for calculating the expected total cost, the recursive equations for the required expected values are first mathematically derived. Then, due to computational complexity, an efficient Monte-Carlo simulation algorithm is designed to calculate these values. The proposed approach is illustrated through practical numerical examples, and the optimal periodic inspection interval over a finite time horizon is obtained via minimization of the expected total cost. Finally, the correctness of the results is validated by sensitivity analysis.
Keywords: Hidden failure, Two-stage revealed failures, Stochastic dependence, Imperfect inspection, Opportunistic inspection.
