Hermite extreme value estimation of non-Gaussian wind load process on a long-span roof structure

This paper presents a combined study of wind tunnel experiment and numerical simulation of the wind-induced pressures on the long-span roof of the Hangzhou East Railway Station Building. Wind tunnel tests were performed on a 1:250-scale rigid model of the station building. Based on the measured pressure data, the third and fourth statistical moments of the pressure processes were evaluated to quantify the non-Gaussian nature of the wind-induced pressures on the station roof. Using the recently reported Hermite moment model, an analytical form of the non-Gaussian peak factor was proposed for a given hardening load process and was verified using numerical integration. The currently available simulation algorithm was revised to generate sample functions of skewed hardening load processes. The simulated pressure data samples provide a basis for the direct statistical analysis of extreme peaks. The peak factors for non-Gaussian wind load effects were estimated by employing various state-of-the-art methods and compared to the mean extreme values of wind pressure data from simulations. © 2014 American Society of Civil Engineers.