A prognostic maintenance policy - effect on component lifetimes

Abstract

Industrial manufacturing systems are becoming more complex; this complexity introduces additional interdependencies between components and systems. To cope with this, new maintenance policies like condition monitoring and prognostics are developed to predict the remaining useful life (RUL) of components. However, decision making based on these predictions is a still underexplored area of maintenance management. The objective of this paper is to quantify the added value of prognostic information (RUL) in maintenance decision making for multi-component systems considering different levels of inter-component dependence (i.e. economic, structural and stochastic). Furthermore, the effect of implementation of the prognostic maintenance policy on the component lifetimes is investigated, as generally in literature the use of prognostics in maintenance scheduling is perceived as to increase component lifetimes. A dynamic prognostic maintenance policy is developed, which takes into account the real component degradation and inter-component dependencies to optimally plan maintenance while minimizing the long-term average maintenance cost per unit time. The added-value of scheduling maintenance actions based on prognostic information is determined by comparing it to two other conventional maintenance policies, these are: age-based preventive maintenance without grouping and age-based preventive maintenance with grouping of maintenance activities. The ability of the prognostic maintenance policy to react to different and changing deterioration patterns and dependencies between all considered components is validated and illustrated by a real life case study on a multi-component manufacturing system. The results show that the developed dynamic prognostic maintenance policy reduces the long-term maintenance costs. Moreover, it is shown that the magnitude of this cost reduction and increase or decrease in component lifetimes depends on the component dependencies.