Deterioration Process Modeling of a Multi-Component Complex System for Condition Based Maintenance

Abstract

A multi-component complex system consists of distri buted components that are connected via a network, e.g., transformers on an e lectricity grid. Loads on the components can be modeled as stochastically varying and may be shared by components connected by the network. Thus, components can be dependent on each other with respect to deterioration, failure, and maintenance initiation. It is anticipated that electricity grid systems will eventually feature sensing and measure ment technologies to enable condition monitoring for components. In response, the purpose of this paper is to present a systemlevel state-space model for condition-based mainten ance (CBM) that incorporates multiple component-level deterioration process models. Component level models are defined based on assumptions that continuous monitoring data and algorithms for condition assessment of a component are available and yield a stochastic dete rioration model of a particular form. Combining these individual deterioration models with a simple model for load sharing, a system state-space model is created and analyzed us ing Monte Carlo Simulation. A scheme for maintenance is then applied to this model and o ptimized for different conditions. The behavior of single-component versus multi-component system models is analyzed and discussed as a possible precursor to more sophistic ated state-space modeling techniques.