基于Pareto分布的风压极值计算方法

Method of estimating extreme wind pressure based on the Pareto distribution

为了利用短时距样本得到具有指定保证率的风压极值,根据风压时程的自相关分析,从短时距风压样本中得到足够的峰值风压样本,通过Pareto分布I型分布拟合峰值样本的高尾部数据,并利用广义极值分布和广义Pareto分布之间的关系对风压的极值做出估计,得到基于Pareto分布的风压极值计算方法。高层建筑表面的围护结构设计取决于结构的表面风压极值的选取。为了获得准确的风压极值,利用高层建筑风洞试验多次独立采样得到的数据,将基于Pareto分布的风压极值计算方法与基于经典极值理论改进的Gumbel法、改进峰值因子法和Sadek-Simiu法进行了比较,从风压极值的期望值和指定保证率的极值两个方面对比可以发现,前者可以得到更为准确的估计。

In order to estimate extreme wind pressure with a determined guaranteed ratio, the Pareto distribution was applied to fit the tail part of peak wind pressures, which wereextracted by using the autocorrelation of a wind pressure time history. The relatithe Pareto distribution and the Gumbel distribution was also used for this estimation. The extreme wind pressure on the surface of high-rise buildings is closely associated with the safety of enclosure structures, so the accuracy of the calculated extreme value is essential. This new approach was applied to a high-rise building model repeatedly tested in wind tunnel experiments. The result show that the new approach delivers better estimations for the expectation and guaranteed ratio of extreme wind pressure than traditional methods such as the Gumbel model, modified Hermite model,and Sadek-Simiu model. Averaged standard deviation and specified warranty of wind pressure coefficient also show that the new approach provides not only moreaccurate but also more robust extreme value than current common methods.