Discrete wavelet transform using Daubechies' wavelet, was applied to analyze a turbulent flow data. For comparison purpose, the turbulent flow signal was phase-randomized, which had the same Fourier power spectrum. The results of the wavelet analysis clearly demonstrated the advantage of the wavelet analysis. The characteristic pattern corresponding to the energy cascade process and strong intermittency of small scales was identified for the turbulent flow data, while the pattern was not discernible for the phase-randomized signal.
Lorentz thresholding methods was applied to the data to separate the attached and detached eddy motion. The decomposed signal by the filtering was found to validate Townsend's hypothesis and Kolmogrov's theory. Also, it was observed that a small fraction of the wavelet coefficients were sufficient to represent the characteristics of attached eddy motion.
