Condition Based Maintenance Optimization of Multi-Equipment Manufacturing Systems by Combining Discrete Event Simulation and Multiobjective Evolutionary Algorithms

Modern industrial engineers are continually faced with the challenge of meeting increasing demands for high quality products while using a reduced amount of resources. Since systems used in the production of goods and deliveries of services constitute the vast portion of capital in most industries, maintenance of such systems is crucial (Oyarbide-Zubillaga, Goti, & Sánchez 2008). Several studies compiled by Mjema (2002) show that maintenance costs represent from 3 to 40 % out of the total product cost (with an average value of a 28%). Within maintenance, the Condition-Based Maintenance (CBM) techniques are very important. Nevertheless, and comparing it to the Preventive Maintenance (PM) optimization problem, relatively few papers related to CBM have been developed: According to Aven (1996), one of the reasons to justify this fact is that CBM models are usually by its nature rather sophisticated compared to the more traditional replacement models. Within this maintenance strategy, Das & Sarkar (1999) distinguish two CBM subtypes, On-Condition Maintenance (OCM) and Condition Monitoring (CMT). OCM is based on periodic inspections, while CMT performs a continuous monitoring on the hardware through instrumentation. Considering the described context, this paper focuses on the problem of CMT optimisation in a manufacturing environment, with the objective of determining the optimal CMT deterioration levels beyond which PM activities should be applied under cost and profit criteria in a multi-equipment system. The initiative considers the interaction of production, work in process material, quality and maintenance aspects. In this work the suitability of discrete event simulation to model or modify complex system models is combined with the aptitude that multiobjective evolutionary algorithms have shown to deal with multiobjective problems to develop a maintenance management and optimisation approach. An application case where the activities applied on a system that produce