Residual life-based importance measures for predictive maintenance decision-making

Abstract

Importance measures have been widely used as meaningful decision-aiding indicators in reliability engineering, risk management and maintenance optimization. However, few importance measures integrates the actual condition (working states or degradation levels) of components that dynamically evolves with time. This work develops a novel time-dependent importance measure defined as the capacity of a component (or group of components) to improve, when it is replaced, the system residual life. The proposed [Formula: see text] measure can help to better prioritize a component or group of components regarding to its improvement ability in the system life time while considering the actual conditions of all components of the system. The originality and complementarity of the proposed measure when compared to existing importance measures is also investigated. The proposed importance measure is then extended to integrate the economic dimension of the maintenance decision, through the maintenance costs, the benefit gained by the maintenance operations and as well as the economic dependence between components. It is finally shown how the proposed [Formula: see text] measure and its extension can “optimally” suggest a component or a group of several components for preventive maintenance decision-making, based on both the technical criterion (residual life of the system) and the economic aspects (benefit and costs). The use and the advantages of the proposed importance measure and its extension are illustrated on a four-component system.