Maintenance of degrading units: new policy and some critical considerations about value of inspections

In this paper, a new maintenance policy is proposed for deteriorating units whose degradation paths are affected by unit-to unit variability. The degradation process is modeled by using a gamma process with random effect. The units are assumed to fail when their degradation level exceeds a (fixed) given threshold. It is assumed that failures are not self-announcing and (consequently) that only an inspection can allow to say with certainty if a unit is failed or not. The maintenance policy is defined by assuming that a unit can be subjected to no more than two inspections. The first inspection is performed at a planned time. Hence, based on the result of the first inspection the unit can be immediately replaced, a second inspection time can be planned, or a future replacement time can be possibly defined. In the latter case, the unit will be replaced at the future replacement time, independently of its state, without performing a second inspection. Differently, when the decision consists in executing a second inspection, at the second inspection time, once again, based on the result of the inspection, the unit can be immediately replaced or its replacement can be postponed to a given successive time, where it will be replaced, independently of its state, without performing a third inspection. After each replacement, the unit is considered as good as new. The results obtained by using the proposed policy are compared with those obtained by using a similar existing (simplified) policy that includes only one inspection and a classical age-based policy, which does not make use of any inspection. The comparison is performed by considering different values of logistic and inspection costs. Obtained results show that depending on the scenario either of the considered policies can be preferred to the others. Some considerations about the value of information gained by performing an inspection close the paper.