Non-stationary analysis on field measured wind speed data of tall buildings

The vector decomposition method was firstly adopted in this paper to process the measured wind speed data atop the super high-rise building Guangzhou West Tower. Three limitations of this widely used statistica analysis method in wind engineering were discussed. In order to solve these limitations and obtaine more accurate fluctuating wind speed time history, the non-stationary characteristics of wind was taken into accoun in this study. Therefore a revised stationary and non-stationary (mainly induced from the EMD method) wind speed model were established respectively for the two orthogonal directions of the fluctuating wind speed obtained from the propeller anemometer. Then the four major indexes of wind characteristic for measured wind speed, i.e. the turbulence intensity, the gust factor, the probability density distribution, and the power spectral density of fluctuating wind speed, were discussed comprehensively by comparing the analyzed results based on the specifications of the two above mentioned wind speed model. The comprehensive compared results showed that nonstationary wind speed model, which was obtained from the EMD method, could describe the true state of wind characteristic of on-site measured wind speed data more accurately.