Nonstationary wind characteristics of wind speed of normal climate at Xijiang bridge site

Based on wind environment monitoring system of Nanning-Guangzhou railway of Xijiang bridge and taking its measured wind field data as the research object, two strong wind samples were chosen. The empirical mode decomposition (EMD) and wavelet analysis methods were applied to analyze nonstationary wind characteristics for these two samples. The nonstationary wind speed models were established by EMD and wavelet analysis methods. Moreover, comparisons were made between traditional wind speed model and nonstationary models. The results show that nonstationary characteristics at Xijiang bridge site are significant and the calculated values of turbulence intensify, gust factor, integral scales by using traditional wind speed model are larger than those by nonstationary models. With respect to probability density distribution of pulsating wind speed, nonstationary models fit well with the assumption of normal distribution in comparison to stationary model. The longitudinal power spectral density is in good agreement at low frequencies by using these two kinds of models, while the calculation values of nonstationary model is less than those of stationary model at high frequencies. Karman spectrum is in good agreement with the results of nonstationary and stationary models, but there exists bigger difference at low frequency. © 2017, Central South University Press. All right reserved.