Gearbox Fault Diagnosis and Prediction Based on Empirical Mode Decomposition Scheme

The empirical mode decomposition (EMD) is a novel method for adaptive analysis of non-linear and non-stationary signals. This paper applies this method to vibration signal analysis for gearbox fault diagnosis. The instantaneous energy density was regard as a feature for gear fault detection. By application of the Hilbert transform on intrinsic mode functions (IMF) mode, the prediction curve based on the average energy values can be derived, which can provide an early warning before final failure. Vibration signals collected from a lathe gearbox with an incipient tooth crack are used in the investigation. The results show that the EMD algorithm and the Hilbert spectrum perform excellently. The instantaneous energy density is shown to obtain high values when defected teeth are engaged and prediction model are proposed.