Improved method of independent storms for extreme value estimation of non-Gaussian wind pressure

In order to determine the design wind load for cladding under a certain guaranteed ratio, a procedure for statistical estimation of extreme wind pressures with small sample size is proposed by improving method of independent storms. The autocorrelation analysis was carried out to extract independent peaks of single wind pressure time history. The weighted least square method was used to obtain the best unbiased estimator of the extreme value. Due to the independence and consistency of the extracted peaks, the new approach could be used to obtain the probabilistic model of extreme wind pressure with small sample size. The approach was applied to two typical large span roofs based on repeated wind tunnel tests. The results show that the new approach usually estimates the extreme pressure more accurately than the traditional method such as peak factor method, modified Hermite model and Sadek-Simiu model. An analytical solution to the quantiles of extreme wind pressure could be obtained with the proposed approach, which shows great potential of design applications.