Characterizing Dependence of Extreme Wind Pressures

Abstract

The prediction of the extreme value of the wind pressure receives significant attention from the wind engineering community. However, the correlation among extreme wind pressures associated with different locations on the building surface is less addressed. In this study, the characterization of the dependence structure of extreme wind pressures and its application to roofing damage estimation are presented. First, three methods to model the dependence structure of extreme wind pressures, including multivariate extreme value theory, copula function, and Nataf transformation, are reviewed. Then, the asymptotic property of the bivariate non-Gaussian random vector will be established and the partial correlation for the data within finite duration is discussed. Furthermore, the dependence structure and asymptotical properties of extreme wind pressure coefficients are investigated based on the very long wind tunnel data. The correlation of extreme pressures considering the randomness of the annual maximum mean wind speed is also addressed. Finally, the application of the correlation of extreme wind pressures on the damage estimation for the asphalt shingle roofing is discussed.