Wind–Temperature Characteristics of a Cable-Stayed Bridge along the Yellow Sea under Superstrong Typhoon Lekima

Abstract

The superstrong typhoon Lekima landed twice along the Yellow Sea of China in 2019; its wind characteristics and impact on bridges’ buffeting performance are still unclear. Based on the measured wind and temperature data of a long-span composite girder cable-stayed bridge along the Yellow Sea, this paper studies the wind characteristics and wind–temperature correlation of typhoon Lekima at the bridge site. First, the non-Gaussian and nonstationary tests of the typhoon are carried out at three stages, namely, before, during, and after the landing of the typhoon. Second, the turbulence intensity and gust factor at the three stages are statistically analyzed, and then the power spectra and the turbulence integral scale are compared by using four evaluation methods. Finally, the joint distribution model of the wind speed and structural temperature at the bottom of the composite girder is constructed, and their parameter estimation and the density functions of five copula functions are calculated, respectively. The results indicate that the wind speed has obvious Gaussian and nonstationary characteristics during the typhoon, while it has non-Gaussian and nonstationary characteristics before and after the typhoon. The gust factor is consistent in characterizing the turbulent characteristics of the fluctuating wind and linearly changes with the turbulence intensity. Besides, the power spectrum is consistent with four classical spectra in the low-frequency region before and during the typhoon, and the autocorrelation index method and power spectrum method are not suitable for calculating its turbulence integral scale. By comparing the square Euclidean distances, root mean square error, and the Kendall and Spearman rank correlation coefficients, the Frank-copula function has the best fitting accuracy for the correction between the wind speed and temperature among the five copula functions with a symmetric U-shaped distribution in thick-tails’ location.