基于广义帕累托分布模型的短期风速资料抗风设计风速计算研究

Design Wind Speed Estimation for the Generalized Pareto Distribution Based on Short-Term Measure Data

在工程抗风设计时常常面临风速资料短缺的问题,基于广义帕累托分布(generalized pareto distribution,GPD)模型,推导短期风速资料进行设计风速取值的方法,推荐适用的阈值选取原则,分析GPD模型计算短期资料设计风速的适用性。结果表明,在使用GPD模型计算短期资料的设计风速时,采用合理的阈值,可得到精度可靠的设计风速供工程设计使用。利用2年的短期观测资料进行设计风速计算时,推荐阈值按"样本均值加Δ倍的标准差"取值,Δ值的选取兼顾样本长度足够长及同时通过"K-S拟合优度检验""以λ为参数的Poisson分布检验"的最大值。在周围空旷、探测环境空旷的地区,采用推荐方法计算短期资料的设计风速时,成果更加可靠。

The shortage of wind speed measured data is often encountered in structural wind resistance design.In this paper,the generalized Pareto distribution (GPD) model is used to derive the method of designed wind speed estimate based on short-term measured data.The method recommends a threshold selection method by an example,and analyzes the applicability of the GPD model.The results show that reliable design wind speed can be calculated precisely based on short-term data by using the GPD model,with a reasonable threshold.And the threshold is according to the mean plus Δtimes the standard deviation,calculating the designed wind speed based on 2-year’short-term measured data.The Δvalue is the maximum one passed "K-S goodness of fit test" and "Poisson distribution test",taken account of the long enough sample length.The results calculated by GPD model are more reliable in the open area than in the area with greater impact of urbanization.