Application of feature extraction methods on a CH-53 swashplate bearing

The purpose of this paper is to compare the performance of features in order to diagnose the swashplate bearing of a CH-53 helicopter by vibration analysis. The aim is to examine which features best discriminate between the faulty and the healthy recordings. Signals were recorded on a full scale test rig. The rig is used to simulate the work environment of the swashplate bearing. Results of commonly used statistical moments are shown to provide poor separation between the faulty and the healthy recordings. Specific, bearing oriented, features based on the energy of the bearing tones, are examined in the order and envelope spectra. The results are shown, presenting a significant performance improvement while indicating the importance of the calculation parameters. This work is an initial step for HUMS implementation on aircraft swashplates, future research is required in order to use these features as health indicators.