• 首页关于本刊投稿须知期刊订阅编委会期刊合作查询检索English

查看全文   查看HTML全文 下载PDF阅读器  

  梁基冠,程印,赵岑.离散Fréchet距离在强震台站场地分类中的应用[J].震灾防御技术,2019,14(4):731-741, DOI:10.11899/zzfy20190404.

离散Fréchet距离在强震台站场地分类中的应用
摘要:    当强震台站场地资料不完整时,所收集到的强震数据因缺乏准确的场地类别信息而难以有效利用。为解决这一问题,本文提出一种基于离散Fréchet距离的强震台站场地分类方法。将获取到的664个KiK-net台站场地按照《建筑抗震设计规范(GB 50011—2010)》进行分类,并构建2个数据集。利用数据集1得到I、II、III类场地标准谱比曲线,并结合离散Fréchet距离对数据集2中的台站进行场地分类,统计分类成功率与误判率。统计结果表明,本文方法能较准确地对场地进行分类,且误判率在可接受范围内。将本文方法分类统计结果与斯皮尔曼秩相关系数法分类的成功率与误判率进行对比,结合本文方法分类后得到的平均谱比曲线,均可表明应用本文方法进行强震台站场地分类具有合理性。
作者单位
梁基冠 西南交通大学, 土木工程学院, 成都 610031 
程印 西南交通大学, 土木工程学院, 成都 610031 
赵岑 西南交通大学, 土木工程学院, 成都 610031 
关  键  词:强震台站  场地分类  离散Fréchet距离  标准谱比曲线  数据集
DOI:10.11899/zzfy20190404
基金项目:国家自然科学基金青年科学基金项目(51708460)
收稿日期:2018-12-19
作者简介:梁基冠,男,生于1993年。硕士研究生。主要从事地震动衰减关系研究。E-mail:leungjg@163.com
通讯作者:程印,男,生于1980年。博士,讲师,硕导。主要从事地震工程相关研究。E-mail:chengyin19830@163.com
Application of the Discrete Fréchet Distance in Site Classifications at Strong Motion Stations
Abstract:      Without complete site information of strong motion stations, the collected data cannot be effectively used due to lack of accurate site category information. In order to overcome this problem, this study proposed a site classification method for strong motion stations based on discrete Fréchet distance, which is a measurement of the similarity between curves. The site conditions at 664 strong motion stations from KiK-net were firstly classified according to the Chinese Code for Seismic Design of Buildings (GB 50011-2010), and then divided into two sets. Standard spectra-ratio curves for site class I, II and III were derived for the sites in dataset1 and there were obvious differences between those three curves. At the same time, the distribution of the natural logarithmic deviation on response spectra periods of those standard curves was also calculated in order to examine the divergence of the dataset. The sites in dataset2 were reclassified based on the method proposed in this paper, and the corresponding success rates and error rates of classifications were calculated. The result shows that not only could the proposed method classify sites with accuracy but also the error rates would be acceptable. The proposed method shows reasonability, by the comparison of the results derived from the proposed method with those from the method of Spearman's rank correlation coefficient, combined with the average spectra ratio curves derived from the proposed method.
Keywords:  Strong motion station  Site classification  Discrete Fréchet distance  Standard spectra ratio curves  Dataset
关闭