Discriminant analysis for helicopter rotor vibration prognostics

Traditionally, preventive maintenance has the drawback of being too expensive because of unnecessary replacements. Corrective maintenance involving replacement or repair after a failure is also expensive because of down time. In this paper, we present discriminant analysis and stepwise discriminant analysis for vibration prognostics. The goal is prediction of failures in drive train components of helicopters, specifically tail rotor drive shaft or engine input shaft. The method has been applied to make predictions of vibrations in the main rotor shaft of Sikorsky helicopter UH-60, using Goodrich Health and Usage Management System (HUMS). The purpose of these predictions is to help enforce rotor vibrations below a threshold to avoid component damages and to reduce aircrew fatigue. The method is particularly useful for prediction of no increase of vibration. This is an important contribution because drive train components of helicopters are normally very robust, thereby the cost of unnecessary preventive maintenance is very high. The predictions made by the model avoid unnecessary maintenance so that costs are lower.