Modelling age replacement policy under multiple time scales and stochastic usage profiles

This paper offers an original methodology to set multi-dimensional maintenance policies for machines whose aging processes require using multiple time scales. It can be considered a generalization of the traditional approach, that usually employs a unique time scale and sets a single age limit to carry out preventive maintenance actions. The methodology also considers situations in which a set of machines are operated using multiple usage profiles. We define usage profile as the relationship between the use of a machine in terms of one main time scale and another scale. In our case study the age of a mining haul-truck component can be best modeled as a combination of operating hours and load cycles since the last overhaul. We compare the results obtained with respect to using a single time scale policy. The comparison shows the importance of the bias in decision making that may arise due to incomplete modelling of the components' aging process.