Optimization of a maintenance scheduling problem through simulation

Abstract

A maintenance scheduling problem was identified at a petrochemical company and it was then proposed to combine the schedules of three linked vessels through the introduction of mini- and major-GO’s. To find the optimal time between a specific vessel’s scheduled maintenance activities, it was decided to use Arena to simulate the maintenance system and observe the influences of proposed changes on the system as a whole. Using the As-Is model as a baseline, the To-Be model was repeatedly executed with different inputs until certain trends were revealed and an optimal solution emerged. It was found that the vessel introduction time into the schedule played a major part when switching over from an old schedule to a new one. The optimal solution’s cycle times are 2 years longer with activities, 3 years apart. This schedule results in decrease of 42% for total time spent on maintenance activities.