基于自组织特征映射的软土震陷动态分级  

Dynamic Classification Research for Subsidence of Foundation in Soft Clay Area Based on Self-organizing Feature Map

作　　者：李波 崔金涛 王威[3] LI Bo;CUI Jin-tao;WANG Wei(Zhongke Jiantong Engineering Technology Co.Ltd.,Beijing 100076,China;Ji'nan Urban and Rural Planning Research Center,Jinan 250099,China;Faculty of Architecture,Civil and Transportation Engineering,Beijing University of Technology,Beijing 100124,China)

机构地区：[1]中科建通工程技术有限公司,北京100076 [2]济南市城乡规划编制研究中心,济南250099 [3]北京工业大学城市建设学部,北京100124

出　　处：《科学技术与工程》2022年第26期11316-11320,共5页Science Technology and Engineering

基　　金：国家重点研发计划(2018YFD1100902-1)。

摘　　要：为解决软土震陷分级中存在的问题,基于以往研究,通过建立以自组织特征映射(self-organizing feature map,SOM)与动态分级控制(dynamic classification method,DT)为基础的耦合式分级模型,为软土震陷分级提供了一种新的思路。通过SOM方法进行仿真分析,筛选出具有关联性的影响因素,根据其相关性强弱,可剔除个别因素,保证各因素之间的独立性。再应用动态分级法对震陷分级控制进行研究,得到了不同分类数下的震陷分级。计算结果与模糊综合评判法进行了对比分析,其结果表明耦合式分析模型具有良好的分级效果。In order to solve the problem in subsidence grading of soft soil,coupling model based on self-organizing feature map(SOM)and dynamic classification method(DT)was established.This method provides a new way for ranking the foundation settlement in soft soil area.Correlation among influences was displayed by SOM simulation analysis.The independence among factors was ensured through weeding out individual factors.Dynamic classification method was applied to study subsidence grading and hierarchy was obtained in different classification data.The performance is compared with fuzzy synthesis assessment.The results show that the coupling model had a good grading effect.

关 键 词：软土 震陷分级 自组织特征映射 动态分级法 

分 类 号：P315[天文地球—地震学]