Effectiveness of Empirical Mode Decomposition Based Features Compared to Kurtosis Based Features for Diagnosis of Pinion Crack Detection in a Helicopter

Canh Ly, Kenneth I. Ranney, Kwok F. Tom, Hiralal C. Khatri, and Harry J. Decker
Submission Type: 
Full Paper
phmc_10_018.pdf511.88 KBAugust 26, 2010 - 10:25am

Features based on empirical mode decomposition (EMD) of measured vibration data under a laboratory controlled environment were developed for a Bell OH-58 helicopter main rotor gearbox. The data were collected at the Vehicle Technology Directorate, NASA, Glenn, Army Research Laboratory. A tooth on the input pinion of the gearbox was notched and run for an extended period at several over-torque conditions to induce a tooth fracture. Vibration data were recorded at regular intervals until a tooth fractured. The EMD features were found to be more sensitive to the gearbox condition than the kurtosis based feature (FM4) and they diagnosed the onset of cracked tooth much earlier.

Publication Control Number: 
Submission Keywords: 
Time Synchronous Averaging (TSA)
Empirical Mode Decomposition
Submitted by: 

follow us

PHM Society on Facebook Follow PHM Society on Twitter PHM Society on LinkedIn PHM Society RSS News Feed