车辆荷载作用下桥梁应变极值估计的阈值选取

发布时间：2018-01-11 07:03

  本文关键词：车辆荷载作用下桥梁应变极值估计的阈值选取　出处：《应用数学和力学》2017年05期 　论文类型：期刊论文

  更多相关文章： 桥梁健康监测 阈值选取方法 应变极值 广义Pareto分布

【摘要】：采用过阈法估计车辆荷载作用下桥梁的应变极值,合理的阈值选取十分关键.阈值选取过大,信息量少,阈值选取过小,广义Pareto分布模型参数估计偏差大.常用的阈值选取方法不能较好地适用于车辆荷载作用下的应变极值估计.基于太平湖大桥车辆荷载作用下1年的应变数据,对拟合结果较好的3种混合分布进行Monte-Carlo(蒙特-卡洛)抽样,对比同一样本基于不同阈值的广义Pareto分布模型的极值估计结果,提出了一种经验式的阈值选取方法.与常用阈值选取方法相比,根据文中方法所得阈值估计的周应变极值分布与实测结果更为接近,估计结果更好.
[Abstract]:Using the over-threshold method to estimate the strain extreme value of bridge under vehicle load, it is very important to select the reasonable threshold value. The threshold selection is too large, the amount of information is little, and the threshold value is too small. The parameter estimation deviation of the generalized Pareto distribution model is large. The commonly used threshold selection method is not suitable for the strain extremum estimation under vehicle load. Based on the response of Taipinghu Bridge under vehicle load for one year. Change the data. Monte-Carlo (Monte-Carlo) sampling is carried out for three mixed distributions with good fitting results, and the extremum estimation results of the generalized Pareto distribution model based on different thresholds of the same sample are compared. In this paper, an empirical threshold selection method is proposed. Compared with the conventional threshold selection method, the estimated maximum value distribution of the circumferential strain based on the proposed method is closer to the measured results and the estimated results are better.
【作者单位】： 合肥工业大学土木与水利工程学院;中南大学土木工程学院;
【基金】：中国博士后科学基金(2015M581982) 安徽省自然科学基金(1708085ME101)~~
【分类号】：U446
【正文快照】： (我刊编委任伟新来稿)引言极值理论指出[5],极值估计通常只与概率密度函数的高尾分布相关,即车辆荷载效应的极值估计应只与重型车有关.广义Pareto分布(generalized Pareto distribution,GPD)模型作为样本右尾数据的拟合模型,广泛运用于洪水高度、风速、气温等的极值估计中,近，

本文编号：1408539