Helicopter Planet Bearing Fault Diagnosis Based on Modulation Signal Bispectrum Analysis

Planet gear bearings are of critical components in the main gearbox of mission critical helicopters. Because of multiple excitation sources, weak cyclisation signatures and vibration transmission modulation (VTM) make it very challenging to diagnose the healthy conditions of planet bearings. This paper proposes a new diagnostic method based on modulation signal bispectrum (MSB) analysis. MSB can effectively suppress random noise by phase alignment and extract deterministic modulation effects at high frequency resolution, which allows the closely distributed and weak bearing signatures to be pull out with high accuracy. It thus can combine VTM effect with bearing signatures to enhance the weak signatures for achieving more reliable detection and diagnosis. Based on datasets from a multiple stage 1300 kW epicyclic gearbox employed in a commercial helicopter, this MSB based method is evaluated under different fault severity and operating conditions. It shows that the amplitudes coupled between fault frequency and epicycle carrier frequency can have more deterministic information regarding to bearing vibrations. Consequently, MSB coherences and magnitudes can jointly give accurate diagnostic results that are consistent with different fault severity and operating conditions, showing outstanding performance of MSB analysis.