A Real-Time Maintenance Policy for Multi-Stage Manufacturing Systems Considering Imperfect Maintenance Effects

Abstract

Machines are subject to deterioration due to usage and aging. Once one machine fails, maintenance action has to be imposed in order to resume its operation. In practice, maintenance actions are not always able to restore the machine as new. Multi-level maintenance is a more realistic scenario. It is important to select a proper maintenance level upon failure since it directly relates to the duration of a maintenance action and the deterioration status after maintenance and therefore determines the overall system performance. In this paper, a real-time maintenance policy is proposed to select an optimal maintenance level aiming at reducing maintenance-related costs. The maintenance cost includes resource cost and production loss due to machine stoppage. The maintenance cost rate, which is the cost per unit time after the maintenance, is used to select the optimal maintenance level. The maintenance effect is modeled with a virtual-age approach. With the available data collected by distributed sensors, a data-driven modeling of serial production lines is established for analyzing production dynamics. The opportunity window of machine stoppage in the stochastic scenario is derived, which is further used to estimate the production loss caused by the machine stoppage due to maintenance. A numerical experiment is conducted to validate the proposed maintenance policy.