A cost-benefit approach for the evaluation of prognostics-updated maintenance strategies in complex dynamic systems

The implementation of maintenance strategies which integrate online condition data has the potential to increase availability and reduce maintenance costs. Prognostics techniques enable the implementation of these strategies through up-to-date remaining useful life estimations. However, a cost-benefit assessment is necessary to verify the scale of potential benefits of condition-based maintenance strategies and prognostics for a given application. The majority of prognostics applications focus on the evaluation of a specific failure mode of an asset. However, industrial systems are comprised of different assets with multiple failure modes, which in turn, work in cooperation to perform a system level function. Besides, these systems include time-dependent events and conditional triggering events which cause further effects on the system. In this context not only are the system-level prognostics predictions challenging, but also the cost-benefit analysis of condition-based maintenance policies. In this work we combine asset prognostics predictions with temporal logic so as to obtain an up-to-date system level health estimation. We use asset level and system level prognostics estimations to evaluate the cost-effectiveness of alternative maintenance policies. The application of the proposed approach enables the adoption of conscious trade-off decisions between alternative maintenance strategies for complex systems. The benefits of the proposed approach are discussed with a case study from the power industry.