Research On Wavelet Base Selection For Vibration Signal Processing

Wavelet analysis, as the typical represent of time-frequency analysis, has been extensively applied to vibration signal analysis. Different wavelet base causing different wavelet spectrum says that variable effects of wavelet analysis can achieve under diverse wavelet base. But the comparison and selection of wavelet base is a difficult problem, which has no system method and theory yet.To compare the different result of denoising by diverse wavelet base, this paper introduces a criterion, which is called reconstruction factor, and gives the calculating formula. It accomplishes the quantificational comparison of each wavelet base used for filtering and denoising by the calculation of reconstruction factor of each situation. It also offers qualitative comparison of wavelet base selection for vibration signal character abstracting.In the end of this dissertation it gives the application example of wavelet such as strain signal denoising, pile foundation integrity testing and impact vibration energy transfer etc, which is analyzed under the selected wavelet base. This dissertation has important value for wavelet theory study, especially offers a creative way to the wavelet base selection. Moreover, it provides certain guiding value for wavelet application in vibration signal processing.