Early Detection of Gear Faults Using Time-Frequency Distribution

The paper presents the results of simulation and experimental tests conducted on toothed gears with programmed damages such as the cracking of a tooth at its base and the chipping of a tooth tip. In the case of the measurements of the acceleration of vibrations in the gear bearing housing, Wigner-Ville (WV) time-frequency analysis proved to be an efficient method to detect impulse disturbances in the broadband vibration spectrum. Three-dimensional functions such as the WV signal of vibration accelerations in a gear with local damages, as presented in the paper, contain much information about the energy distribution of vibrations which were caused by local damages. The additional symptom confirming susceptibility to local damage is the kurtosis coefficient determined in a movable time window with a properly selected width. The tests presented in the paper confirm that the application of WV analysis together with classical methods of signal analysis can increase the efficiency of local damage diagnostics.