Early detection of gear failure by vibration analysis i. calculation of the time-frequency distribution

The application of the spectrogram to the calculation of the time-frequency distribution of a gear vibration signal is examined. The spectrogram represents the energy distribution of the signal over the frequency domain at every instant of time. The selection of the window function is discussed, and it is shown that the Gaussian function is well suited to the calculation of the energy distribution, giving a representation which is free from ripple and easy to interpret. The application of the energy distribution is demonstrated using numerically-generated test signals and the experimentally-measured time domain average of the gear meshing vibration from an industrial gearbox. The results suggest that the spectrogram may provide a powerful tool for the early detection of local gear damage. Copyright 1993, 1999 Academic Press