考虑参数更新的大跨桥梁温差极值分布估计方法研究

发布时间：2018-05-07 12:20

本文选题：桥梁工程 + 结构健康监测　；参考：《工程力学》2017年03期

【摘要】：为了建立可靠的大跨桥梁全寿命温差极值分布模型,提出采用广义帕累托分布(Generalized Pareto Distribution,GPD)对超阈值温差的统计特征进行描述,并给出了超阈值温差样本相关性的去除方法和最优阈值的确定方法。为了融合温差分布的先期经验信息和不断递增的温差监测样本,建立了考虑参数更新的贝叶斯估计方法,利用Gibbs抽样对贝叶斯后验分布进行计算,进而得到准确的基于广义帕累托分布的温差极值分布模型。最后利用九堡大桥长期监测温差数据进行了验证。研究结果表明,广义帕累托分布能够对超阈值温差样本的尾部统计特征进行准确描述,提出的考虑参数更新的温差极值分布贝叶斯估计方法能够对广义帕累托分布的参数进行可靠估计,估计的统计模型比极大似然估计计算的结果更接近真实情况。研究结果可为大跨桥梁温差特性分析提供参考。
[Abstract]:In order to establish a reliable maximum value distribution model of temperature difference for the whole life of long-span bridges, the generalized Pareto distribution and generalized Pareto distribution are used to describe the statistical characteristics of the temperature difference over threshold. The method of removing the correlation of temperature difference samples and the method of determining the optimal threshold are given. In order to integrate the prior empirical information of temperature difference distribution with increasing samples of temperature difference monitoring, a Bayesian estimation method considering parameter updating was established, and the Bayesian posterior distribution was calculated by Gibbs sampling. An accurate extreme temperature distribution model based on the generalized Pareto distribution is obtained. Finally, the long-term monitoring temperature difference data of Jiubao Bridge are used to verify. The results show that the generalized Pareto distribution can accurately describe the tail statistical characteristics of over-threshold temperature difference samples. The Bayesian estimation method of temperature difference extremum distribution considering parameter updating can estimate the parameters of generalized Pareto distribution reliably. The estimated statistical model is closer to the real situation than the maximum likelihood estimation. The results can provide a reference for the analysis of temperature difference characteristics of long span bridges.
【作者单位】：河海大学土木与交通学院;
【基金】：国家自然科学基金项目(51678218) 中国博士后科学基金项目(2014M560387)
【分类号】：U446

【相似文献】