Scheduling aircraft engine maintenance: Modeling and optimization.

摘要

The engine is one of the most critical parts of an aircraft. It provides thrust, electric and hydraulic power needed for an aircraft operation. Failure of an engine can lead to a crash. Clearly, the components in an engine must be highly reliable. Since components are subject to wear, this reliability must be achieved through frequent inspection, repair, and replacement of components. Obviously, this is costly. Finding an appropriate balance between safety and cost is a fundamental problem.;This dissertation addresses some specific engine-related maintenance problems. First, we consider the compressor and the turbine, which are two of the most safety critical components in an engine, in isolation. Because these components have low accessibilities, maintenance of the compressor and turbine is critical from both safety and cost standpoints. We model the compressor problem as a partially observable cumulative shock model and the turbine blade replacement problem as a completely observable cumulative shock model, and demonstrate that simple structured policies are optimal.;Second, we consider the interactions between components with regard to maintenance. Since aircraft are very expensive, the cost due to aircraft downtime is much greater than the downtime cost caused by regular machine maintenance (e.g., in production machinery). Also, in general, whenever maintenance is performed on an engine component, the engine must be removed from the aircraft, which is costly and time consuming. This makes it potentially cost-effective to perform maintenance on several engine components at the same time. However, because the literature on the multi-component joint maintenance problem indicates that, in general, there is no simple optimal structural policy. Therefore, we develop three heuristic approaches that are well-suited to this problem. Among these, the base interval approach, which mandates the replacement of each component to be multiple of a certain base interval, is shown to be the best from both cost and practicality points of view. This result validates the ABC maintenance approach currently practiced in airline industry and provides a computational procedure for indentifying the optimal base interval policy.