非高斯脉动风压的模拟研究

对于风荷载数值模拟，大多数研究都是假定风荷载为平稳高斯随机过程。然而，在分离流作用的一些局部重要区域，例如建筑物屋盖边缘、屋面转角等，风荷载表现出强烈的非高斯特性。为了能够有效地模拟非高斯脉动风压，本文提出一种模拟非高斯脉动风压的框架。首先，根据风速与风压间的关系，严格推导出脉动风压功率谱密度函数。然后，基于导出的功率谱密度函数、Johnson转换系统和数字滤波理论，提出一种能够快速而有效的生成指定偏度、峰度和功率谱非高斯脉动风压的方法。最后，通过一个一维单变量非高斯脉动风压的模拟算例对该方法的可行性和正确性进行验证。数值结果表明：模拟生成的单样本非高斯脉动风压的统计参数例如偏度和峰度与目标偏度和峰度非常吻合，而且模拟功率谱与目标功率谱两者也吻合得很好。

Simulation of non-Guassian fluctuating wind pressure

In the simulation of wind loads, it is mostly hypothesized that wind loads are stationary Gaussian stochastic processes. But, inasmuch as the wind pressure fluctuations measured at corners as well as other separation flow regions of roofs of buildings are found to be highly non-Gaussian, the Gaussian assumption is not appropriate for the generation of the wind pressure time history in these local regions. In order to simulate the non-Gaussian fluctuating wind pressure, the following framework is proposed in the present paper. Firstly, the power spectral density (PSD) function of fluctuating wind pressure is deduced rigorously by resorting to the relationship between the wind pressure and velocity. Then, an efficient yet practical method by which the fluctuating wind pressure may be reproduced with the designated skewness and kurtosis is presented based on the derived PSD function, Johnson translator system, and digital filter. Finally, in order to corroborate the feasibility and correctness of the present methodology, a numerical example of simulating one-dimensional uni-variate non-Guassian fluctuating wind pressure is taken into consideration. Numerical results indicate that the statistical parameters, such as the skewness and kurtosis, of the simulated fluctuating wind pressure are in remarkably agreement with their targets. Likewise, the PSDs of the simulated records and target show considerably good agreement with each other.