Sampling technique in wavelet analysis of vibrating signals of rotating machinery
Qin Shu-ren · Proceedings of the 17th International Modal Analysis Conference · 1999
A main feature of wavelet analysis is localizable characteristic in time-domain and frequency-domain. However, localizable characteristic is formed from time-frequency windows with elasticity. Sampling interval of wavelet transform can be automatically adjusted in the second wavelet sampling, therefore wavelet transform can focus to arbitrary details of observed signal Sample is the important task of wavelet analysis. Sample of wavelet analysis and sample of Fourier analysis have great difference. Sample interval of Fourier analysis is a constant in time and frequency domains. Sampling intervals of wavelet analysis are gradually becoming fine with increasing frequency. On the other hand, after the first sampling of the signals (gotten by A/D converter) and sampling of wavelet base, wavelet transform needs the second sample of the signals and the wavelet base. In signal analysis Shannon theorem on sampling must be satisfied. In this paper, sampling principle and technology of orthogonal wavelet transform of signals are deeply researched. Moreover, edge effect of wavelet analysis of signals is discussed. In this paper, the wavelet analysis of singular signal and vibrating of rotating machinery is discussed, and some application examples are given.