Using wavelet transforms in gearbox vibration monitoring data
- Univ. of Hertfordshire, Hatfield (United Kingdom)
- Brunel Univ., Uxbridge (United Kingdom)
Gearbox vibration signals are composed of transient changes and they are non-stationary (i.e. their spectra change with time). In a mechanical system consisting of various rotating parts, different components contribute to the overall vibration signals at different times and with different levels. Thus, if a signal analysis method used as an alternative to the standard methods can localize the information on the time axis, it can be possible to determine the approximate location of the fault or damage. Hence further techniques have to be developed such that the localized information in the time domain can be mapped to the frequency domain. Wavelet transforms which is an adjustable windowed analysis method have been used to compensate for the inadequate information provided by other signal analysis methods used before. In order to get more meaningful interpretation of the data mean-square wavelet contour maps were used. These maps show how the mean-square value of the signal is distributed between wavelet levels. Mesh diagrams are then generated for those maps for two different types of wavelet transforms which are dilation wavelets and harmonic wavelets.
- OSTI ID:
- 501736
- Report Number(s):
- CONF-961239--
- Country of Publication:
- United States
- Language:
- English
Similar Records
Wavelet transform techniques and signal analysis
Wavelet transform techniques and signal analysis