An integrated time domain averaging scheme for gearbox diagnosis

The vibration signal of a gearbox carries the information of the running parts including gears. The early fault can be detected by the analysis of the vibration signal collected from the gearbox. Traditionally, time domain averaging (TDA) has been used to extract the waveform with respect to each gear from the noisy vibration signal for the purpose of fault diagnosis. However, this method may not work if it is not used properly. This paper gives a deep investigation on how to use TDA properly to get the fault signature for gearbox prognosis. In this application, a damaged gear is recognized, however the fault signature can not be captured by other methods such as FFT and wavelet transform.