Non-Gaussian properties and their effects on extreme values of wind pressure on the roof of long-span structures

This study investigates non-Gaussian properties and their effects on peak factors and extreme values of wind pressure on the irregular roof of a long-span structure. The translation process theory to determine the peak factor of a non-Gaussian process is reviewed. A wind tunnel test is then conducted on the rigid model of the satellite concourse to measure wind pressure on the roof. Spatial and statistical distribution of non-Gaussian properties of wind pressure on the whole roof is illustrated. Three regions on the roof having strong non-Gaussian properties are identified based on the spatial distribution. Peak factors based on the translation process theory are then calculated. The spatial distribution of all the peak factors on the roof is outlined and the regions on the roof having large peak factors are identified. Results show that a peak factor of 2.5 or peak factor gpeak,G based on Gaussian process is far smaller than that determined by the translation process method and hence usually underestimates extreme values of wind pressure by 30%–40%. Therefore, it is highly recommended to consider non-Gaussian properties when evaluating the extreme value of wind pressure on similar long-span structures, especially for the identified sensitive regions.