MODELING A WRONG MAINTENANCE POLICY

The behavior of a single unit system, which is maintained according to the preventive policy, despite it fails with a constant failure rate, is focused. A discrete event simulation (DES) enables to compare the expected number of failures which are obtained by the simulation of alternative maintenance scenarios: the corrective and preventive ones. The results comparison shows, at a first glance, that the preventive maintenance has to be preferred. A deeper analysis, based on the information content which can be registered at each step of the simulation process, on the Skellam function properties and on the Small Number Law, helps to clarify this strange behavior. The real aim of modeling such a system would be to refuse the constant failure rate as an operations statement but only as a missing-information maintenance state. The concept of random failures is valid in a statistical sense but it should not discourage efforts to predict and prevent individual failure by all means open to us. The simulation model enables to find more comprehensive information content about the behavior of stochastic single unit maintenance.