Helicopter Main Gearbox Planetary Bearing Fault Diagnosis using Vibration Signal Processing Techniques

Helicopters are extensively employed as versatile assets for military, transportation, rescue and many other utilities. A healthy and functional Helicopter Main Gearbox (MGB) is critical to ascertain the helicopter’s system reliability and flight airworthiness, i.e. the suitability for safe flight. Therefore, it is of great importance to monitor the health status of the MGB. Currently, Health and Usage Monitoring System (HUMS), has been deployed on all medium and large size civil helicopters to perform MGB health status monitoring in United Kingdom. Nevertheless, HUMS has shown insensitivities
and a lack of accuracy to detect planetary bearings-related defects, resulting in unfortunate accidents. Therefore, the successful diagnosis of planetary bearing defects in MGB could contribute profoundly to enhance sensitivity of HUMS against such defect type, thus improving helicopter flight safety, and reducing the overall maintenance costs.
This research aims at investigating the diagnosis of planetary bearing faults inside MGB using advanced signal processing techniques, providing diagnostic information that is more accurate and indicative against incipient planetary bearing faults. To fulfil the requirements, experimental work was undertaken on a commercial helicopter MGB to acquire invaluable vibration data. The MGB was operated under various load, speed and fault severities conditions. Diagnosis of the seeded planetary bearing faults was then successfully performed by evaluating and implementing various frequency domain processing techniques. Finally, further evaluation was conducted using another MGB dataset collected from a CH-46E Aft gearbox. The results of this study have shown that applying various frequency domain signal processing techniques can effectively detect incipient planetary bearing faults. The main contributions of this research include acquiring data from a full scale helicopter main gearbox, proposing and evaluating a non-deterministic weak signature analysis scheme for MGB planetary bearings fault detection, and demonstrating using MGB carrier induced sidebands as a novel spectral feature for planetary bearing diagnosis.