Abstract

The considerable increase in the complexity associated with the formulation of maintenance plans has enabled the development of new techniques to bring maintenance scheduling optimization models to more realistic environments. In this sense, a previous optimization model was proposed considering the use of time windows for the formation of grouping schemes under an opportunistic strategy for maintenance activities considering non-negligible execution times, thus offering the possibility of analysing scenarios with limited resources. This article proposes a risk analysis based on the failure probability of each component involved in the maintenance scheduling optimization model, which has the particularity of enabling a greater number of combinations of grouped PM activities. Moreover, it seeks to identify the general behaviour of the optimization model against different scenarios of periodicities and execution times of each maintenance activity. The proposed optimization model is formulated under a mixed integer linear programming (MILP) paradigm and its objective function seeks to minimize the unavailability of the system associated with the execution times of the activities developed, generating different experimental cases, and varying the start time scheduling under a tolerance factor from 0% up to a maximum of 25% for advance or delay. Results show in contrast with the base optimization model, an 8% less unavailability when the tolerance factor is 10%. Finally, it was possible to quantify the risk present in each maintenance schedule, at the same time a behaviour towards advancing PM activities is evidenced by the optimization model proposed over the delay.