Optimization of preventive maintenance of multi-component mechanical systems by genetic algorithms: case of the ammonia unit turbo-compressors – FERTIAL

Abstract The supervision of the maintenance of industrial facilities, especially the expenses related to the application of preventive strategies, is extremely fascinating because of the rising relevance of this service in the production chains. This article represents a model of periodic preventive maintenance policy (fixed period/ variable period) based on the concept of reducing the age of the components of a system. So that, three maintenance activities can be applied to components, simple preventive maintenance action, preventive revisions and preventive replacement.The combination of optimum activities entails determining the preventive maintenance actions and intervals necessary for them on each stage using genetic algorithms. The objective is therefore to determine the activities and the optimal interval at each preventive maintenance step, increasing the lifetime of a system subjected to degradation, considering the maintenance cost for the three types of activities and the minimum repair costs for each preventive maintenance step.In this work, we applied the optimization technique by the genetic algorithm in order not only to minimize the cost of one of the preventive maintenance policies, but also to optimize the activities of a multi-component system. The optimal activities and intervention interval that maximize the life of the system in unit of cost at each preventive maintenance stage can be determined using genetic algorithms, thus, preventive maintenance planning will be determined from one stage to the next until the life in units of maintenance cost is less than its life at disposal.