A competing risk model for dependent and imperfect condition–based preventive and corrective maintenances

Abstract

This article develops a model for dependent and imperfect condition–based preventive and corrective maintenance actions. The approach is based on the combination of the intensity proportional repair alert, a competing risks-based model and the generalized renewal process. Typically, intensity proportional repair alert can identify either how preventive actions may modify the distribution of the time between critical failures or how corrective events may change the frequency of preventive maintenances, but this method fails to analyze the effectiveness of the maintenance actions because they are treated as being perfect. On the other hand, generalized renewal process is able to capture the quality of maintenance, classifying it as perfect, minimal or imperfect depending on the value of a rejuvenation parameter. However, generalized renewal process cannot distinguish how different types of maintenance influence each other as intensity proportional repair alert does. Therefore, the intensity proportional repair alert–generalized renewal process hybrid approach is proposed here to fill this gap. This article also develops the maximum likelihood estimators for the proposed model as well as a Monte Carlo–based algorithm to estimate the expected number of preventive and corrective maintenances over time. The proposed model is validated through two example applications for which the intensity proportional repair alert–generalized renewal process model results show close agreement with the failure datasets.