基于核拟合优度统计法的车道荷载极值模型中的阈值选取  

Threshold Selection in Lane Load Extreme Value Model Based on Kernel Goodness of Fit Statistics

作　　者：毛明春 MAO Ming-chun(Weifang Municipal Engineering Design&Research Institute Co.,Ltd.,Weifang 261000,China)

机构地区：[1]潍坊市市政工程设计研究院有限公司,山东潍坊261000

出　　处：《中国市政工程》2022年第2期66-71,110,111,共8页China Municipal Engineering

摘　　要：基于某斜拉桥一年的动态称重系统(WIM)数据,研究车道荷载的概率分布。研究表明,由于小型车辆占比很高,所以车道荷载的概率密度分布是典型的拖尾分布。为研究车重尾部数据,通过核拟合优度统计量法得到阈值。确定阈值后,采用广义帕累托分布(GPD)拟合车重的尾部数据,采用K-S检验方法进行分布类型的拟合检验。检验结果表明,各车道车重尾部数据适应GPD分布,为类似车辆荷载建模提供参考。Based on the one-year dynamic weighing system(WIM)data of a cable-stayed bridge,the probability distribution of lane load is studied.The research shows that due to the high proportion of small vehicles,the probability density distribution of lane load is a typical trailing distribution.To study the tail data of vehicle weight,the threshold is obtained by kernel goodness of fit statistics.After determining the threshold,the generalized Pareto distribution(GPD)is used to fit the tail data of vehicle weight,and the K-S test method is used to fit the distribution type.The test results show that the vehicle weight tail data of each lane does not reject the GPD distribution,which provides a reference for similar vehicle load modeling.

关 键 词：车辆荷载 过阈值模型 核拟合优度统计量法 阈值 

分 类 号：U441.2[建筑科学—桥梁与隧道工程]