Comparison of several non-stationary methods for analysis of time history of non-stationary wind pressure

The method for non-stationary analysis is a major issue in the field of signal analysis. Taking the signal of non-stationary wind pressure and lift coefficient obtained from wind tunnel experiment, a time frequency analysis was conducted by using short-time Fourier transform, Wigner-Ville distribution, wavelet transform, and Hilbert-Huang transform(HHT) separately. The results show that the methods of short-time Fourier transform, Wigner-Ville distribution and wavelet transform are based on Fourier transform, therefore, the spectrum characteristics obtained by these methods are roughly the same with different accuracy. Based on the method of HHT, the meaningful instantaneous frequency and an exact expression of frequency which changes with time were obtained. The signal was expressed as the energy distribution in the time frequency plane as Hilbert spectral. This method can be applied to analyzing a large amount of nonlinear and non-stationary signals whose frequencies change with time, and complex signal can be decomposed into a number of intrinsic mode functions from high- to low- frequencies. However, no matter what kind of analytical method is used, it can be seen that the signal energy is mostly dispersed in the low frequency with small proportion in the high-frequency. Therefore, one of these methods can be used to extract the controlling low-frequency component and to express it through regression as a simple wind-signal, and then the signal can be applied to the structure to study the wind characteristics.