基于聚类分析的我国工程场地分类方案优化研究  

Research on optimization of site classification scheme in Chinese seismic code based on clustering analysis

作　　者：任叶飞[1,2,3] 刘也 张鹏 冀昆[4] 王宏伟 温瑞智 谢俊举[5] REN Yefei;LIU Ye;ZHANG Peng;JI Kun;WANG Hongwei;WEN Ruizhi;XIE Junju(Institute of Engineering Mechanics,China Earthquake Administration,Harbin 150080,China;Key Laboratory of Earthquake Engineering and Engineering Vibration,China Earthquake Administration,Harbin 150080,China;Key Laboratory of Earthquake Disaster Mitigation,Ministry of Emergency Management,Harbin 150080,China;School of Civil and Transportation Engineering,Hohai University,Nanjing 210098,China;Institute of Geophysics,China Earthquake Administration,Beijing 100081,China)

机构地区：[1]中国地震局工程力学研究所,黑龙江哈尔滨150080 [2]中国地震局地震工程与工程振动重点实验室,黑龙江哈尔滨150080 [3]应急管理部地震灾害防治重点实验室,黑龙江哈尔滨150080 [4]河海大学土木与交通学院,江苏南京210098 [5]中国地震局地球物理研究所,北京100081

出　　处：《建筑结构学报》2023年第11期226-235,共10页Journal of Building Structures

基　　金：国家重点研发计划项目(2018YFE0109800);国家自然科学基金项目(51878632);黑龙江省自然科学基金优秀青年项目(YQ2019E036);黑龙江省头雁行动计划。

摘　　要：针对我国现行抗震设计规范中Ⅱ类场地定义范围是否过于宽泛的问题,根据日本KiK-net台网的强震动记录和台站钻孔资料,基于地震动预测残差分析方法计算台站的真实场地放大因子,引入K-Means聚类算法根据台站真实场地放大因子对Ⅱ类场地进行聚类分析,给出建议的场地分类优化方案。研究结果表明:Ⅱ类场地的地震动场地放大因子离散性较大,存在与覆盖层厚度相关的明显分簇现象,验证了Ⅱ类场地定义范围存在过于宽泛的问题,也为聚类分析提供了先决条件;当Ⅱ类场地的台站被划分为2个簇类时,聚类结果最优,且两个簇类的场地放大因子曲线差异明显、覆盖层厚度总体上呈分簇现象,验证了聚类结果的可靠性;建议以覆盖层厚度20 m作为分界指标将Ⅱ类场地细分为Ⅱa和Ⅱb两个子类,两个子类的场地放大因子具有较好的类间差异度与类内聚集度。It is debated widely whether the scope of site classⅡin current Chinese seismic codes is too broad or not.In this study,strong motion records and borehole data from KiK-net seismograph networks were collected to calculate the site amplification factors using the method of residual analysis of ground motion model.Then a K-means clustering algorithm was introduced to implement clustering analysis on stations of site classⅡby setting the site amplification factors as explanatory variables.An optimized site classification scheme was suggested finally.The results show that:the site amplification factors of ClassⅡsites show great variability and clustered obviously with respect to soil covering depth,implying broad scope of definition of site classⅡand providing a prerequisite for cluster analysis;it is recognized that the optimum number of clusters is two in order to achieve the best clustering site classⅡstations;the site amplification factor curves of two clusters are significantly different and their soil covering depths are clustered obviously,hence validating the reliable results of K-means clustering;it is recommended that the site classⅡcould be subdivided into two sub-classesⅡa andⅡb considering the 20 m of soil covering depth as a threshold value;both sub-classes reflect good performance of inter-class aggregation and intra-class variation of site amplification factors.

关 键 词：场地分类 场地放大因子 抗震设计规范 K-MEANS聚类算法 覆盖层厚度 

分 类 号：P315.9[天文地球—地震学]