Gearbox fault diagnosis under non-stationary conditions with independent angular re-sampling technique applied to vibration and sound emission signals

Abstract

An analysis of gearbox vibration signals is almost always the default choice when diagnosing the condition of a gearbox because of the rich information contained in the vibration signals and their ease of measurement. At times, however, the gearbox vibration signal may not be available and the sound emission signal may serve as an alternative to diagnose the condition of the gearbox. Gearbox vibration and sound emission signals are mostly non-stationary owing to uncertainties associated with the drive and load mechanisms. The signals acquired from the gearbox are then required to be converted into stationary signals for further analysis. In the present work, the independent angular re-sampling (IAR) technique is employed to convert non-stationary vibration signals (measured in two mutually perpendicular directions) and sound emission signals into quasi-stationary signals in the angular domain. The resulting angular domain averaged (ADA) signals for each gear health condition are then decomposed with continuous wavelet transform and continuous wavelet coefficients (CWCs) fed directly to a back propagation neural network with the objective of diagnosing the condition of the gearbox. Promising results are obtained when the sound emission signals are analyzed to diagnose the condition of the gearbox.