Optimal task ordering for troubleshooting systems faults

Automated troubleshooting of system faults is an essential element of modern aerospace equipment. The increased efficiency and accuracy helps in dramatically lowering maintenance costs. The Prognostic Health Management (PHM) group at Pratt & Whitney is responsible for developing automated systems for fault detection, isolation, and accommodation for the Joint Strike Fighter (JSF) propulsion system. A fundamental question that arises in this context is the following: Given a list of suspected components that have been identified a priori as possible causes for failure symptom(s), what is the optimal troubleshooting task assignment strategy? This paper introduces an approach to optimal task ordering. We show that the correct strategy is to order the tasks based on an easily calculated metric - which we call the mean utility function - that takes into consideration the mean troubleshooting time, or cost, or a combination of the two, depending on what is considered to be most critical. A mathematical proof is given for this. The approach shown in the paper can also be applied, as a troubleshooting strategy, for any other machinery health management system