Gear fault detection under time-varying rotating speed via joint application of multiscale chirplet path pursuit and multiscale morphology analysis

Abstract

This article reports a new method for gear fault detection under time-varying rotating speed. This method is based on the chirplet path pursuit and multiscale morphology analysis. The instantaneous rotating speed is extracted from the gear vibration signal using the multiscale chirplet path pursuit algorithm. According to the extracted rotation speed, the gear vibration signal is resampled at constant angle increment and as such the nonstationary signal is converted into a stationary signal. The fault-induced impulsive features can then be extracted from the resampled signal via the multiscale morphology analysis, followed by the spectrum analysis to reveal the fault characteristic frequency. Because of the low correlations between the noise and chirplet functions, the rotational speed can be extracted effectively even when the signal-to-noise ratio of the vibration signal is relatively low. In addition, the noise effect can be further suppressed by averaging the results of morphology analyses of all the scales. Therefore, the proposed approach has a good antinoise ability and is suitable for gear fault detection under time-varying rotational speed. The performance of the method has been validated by both simulation and experimental data.