基于Copula函数的杭州地区多风向极值风速估计

Multi-directional extreme wind speed estimation in Hangzhou using Copula functions

为了合理确定建设场地的结构抗风设计风速,提出多风向极值风速估计方法.以杭州地区日极值风速风向为统计样本,采用核密度-广义帕累托混合模型来描述各风向下的极值风速边缘分布,提高了Up-crossings法估计极值风速的准确性.基于严格选择的Copula函数建立杭州地区风速风向联合分布模型.采用该联合分布模型计算得到了不同重现期下考虑风向相关性的多风向极值风速,并与不考虑风向相关性的极值风速估计值进行对比.结果表明,具有给定重现期的设计风速估计应考虑风向相关性的影响.在不考虑风向相关性的情况下,容易给出偏于风险的极值风速估计值.

A multi-directional extreme wind speed estimation method was proposed in order to determine the wind-resistant design wind speed for a construction site.The Kernel Density-Generalized Pareto Model was introduced based on the daily maximum wind speed data series in Hangzhou area to describe the marginal distributions of the multi-directional extreme wind speed such that the accuracy of the Upcrossings method was improved for estimating the extreme wind speed.The joint probability distribution model of the multi-directional extreme wind speeds was established based on the selected Copula functions.The extreme wind speeds associated with various return periods were estimated for different wind directions by considering the multi-directional correlations,and compared with those results without considering the correlation.Results show that the estimation of the design wind speed should consider the influence of the wind direction correlation.It will be in the risky side in estimating the extreme wind speeds without considering the multi-directional correlations.