System-Level Maintenance Policies via Stochastic Petri nets with Aging Tokens

This paper describes modeling of system-level maintenance policies using stochastic Petri nets (SPN) with aging tokens. It demonstrates the ability of SPN to capture complex dynamic interaction among various factors affecting maintenance effectiveness both at the component and system level in a visual yet rigorous fashion. In particular, flexible modeling of degradation processes at the component level is combined with capturing the effects of external constraints imposed by system-level considerations. The paper considers two examples of SPN modeling that demonstrate basic building blocks for evaluating various operation and maintenance policies by means of system-level metrics such as total operating costs. The first example addresses opportunistic repair and provides a comparison with optimal replacement policies at the random and fixed intervals. The second example deals with condition-based maintenance based on imperfect inspection. Quantification of the benefits and the drawbacks of the corresponding maintenance policies is provided