Gearbox Vibration Analysis Using a Spectrogram and Power Spectrum Approach

Abstract

Vibration analysis is essential in rotating machinery fault diagnosis. As a vibration contains the dynamic information of a machine, improvement based on analysis has an effective role in predictive and preventive maintenance. In the present paper, the short Fourier transform is applied to determine the frequency variation of a gearbox signal with time due to different loads and driver speeds. In addition, the power spectral density (PSD) is used to represent the randomness of the signal since many frequencies occur simultaneously. The gearbox health condition is measured, and signal fault is simulated as tooth breakage for five cases: 0% (healthy), 25%, 50%, 75%, and 100% (complete tooth breakage). The obtained results proved that it is more powerful to use both spectrograms and PSD for gearbox fault diagnosis. This method is also improved with the ability to distinguish gearbox vibration signals for anomaly detection.