The Study Of Peak Factor For Non-Gaussian Wind Pressure On Long-Span Roofs

With the development of society and the progress of science and technology, constructions become much taller and their structures tend to be much softer. The effect of wind has been very important. The impact of wind load has gradually changed from "be considered" to "must be considered", which means wind load has become one of control loads for structure design.In terms of long-span structure, owing to its huge size and characteristic bluff body shape, when wind is baffled, it will have separation bubbles, shedding vortex and reattachment and all of them are called signature turbulence. Air flow separate and eddy off the edges, corners, ridges of roofs and that cause conical vortex and columnar vortex swirling on roofs. Time history of wind pressure in these areas often performs wind peak pressure events. And high frequency peak value has the same frequency as that of vortex shedding, which is the main reason of the destruction of roof components.In most cases, current load code cannot meet the design requirements of long-span roofs. Instead, model wind tunnel test is a good way to get wind load characteristics of the long-span roofs. Peak factor used in wind tunnel test is based on the theory of Gaussian stationary random process combined with experience of wind engineering. Because of the strong signature turbulence on long-span roofs, wind load show significant Non-Gaussian characteristics on roof edge. For these areas, using Gaussian peak factor to determine the extreme wind load will probably lead to unsafe design.Based on above background, the core of this paper is to study some methods of determining peak factor of Non-Gaussian wind pressure. The main contents in this paper are shown as following:(1) Introduce the method for determining peak factor of Gaussian wind pressure.(2) Focus on the properties and applicable range of Gram-Charlier series and moment-based Hermite model. Deduce a set of explicit expression of third-order coefficient of moment-based Hermite model based on previous studies.(3) Apply Gram-Charlier series and moment-based Hermite model to fitting probability density of Non-Gaussian wind pressure. Furthermore, using them to explain principle and methods for determining peak factor of Non-Gaussian wind pressure.(4) Calculate peak factors of two typical long-span roofs by using the methods mentioned above and compare the results to get a conclusion that the third-order coefficients of moment-based Hermite model deduced in this paper is more appropriate to calculate peak factor of Non-Gaussian wind pressure.