采用小样本统计理论的隧道围岩分类  

Classification of Tunnel Surrounding Rock Based on Small Sample Statistical Theory

作　　者：苏永华[1] 马宁[1] 胡检[1] 

机构地区：[1]湖南大学土木工程学院,湖南长沙410082

出　　处：《公路交通科技》2010年第8期66-69,80,共5页Journal of Highway and Transportation Research and Development

基　　金：GDUE开放基金项目(SKLGDUEK0915);湖南省自然科学基金项目(09JJ3113);湖南省交通厅科技项目(200717)

摘　　要：隧道围岩级别判断存在较多的不确定影响因素,特别是各影响因素的参数信息和样本资料有限,给判断工作造成了很大的困难。利用专为小样本统计分析而提出的支持向量机方法,在分析围岩级别划分的基础上,选取隧道围岩级别划分需要考虑的9个关键因素,将这9个因素作为输入参数,同时将围岩划分为5个等级作为输出参数,建立了围岩级别判断的小样本统计模型。利用该模型对二郎山隧道的围岩级别进行测试,并与ART1神经网络和BP神经网络的结果进行对比,表明将基于小样本统计的支持向量机理论用于围岩级别判断是可行的,并且具有很好的精度。There are many uncertainties in grade distinguishing of surrounding rock,especially limitation of parameter information and sample data of various influencing factors,which has caused difficulty for grade distinguishing of surrounding rock.After analyzing the foundational theory of classification of surrounding rock,by using the method of support vector machine proposed for small sample statistics,9 key influencing factors for tunnel surrounding rock classification were selected as input parameters,and the surrounding rock was classified into 5 grades as output parameters.Then the small sample statistical model for distinguishing the grade of surrounding rock was established to test the surrounding rock grade of Erlang tunnel.The test results were compared with those worked by the methods of ART1 neural network and BP neural network.It indicates that distinguishing grade of surrounding rock with the method of support vector machine based on small sample statistics is feasible and accurate.

关 键 词：隧道工程 围岩 判别模型 级别判断 小样本统计理论 

分 类 号：TU451.2[建筑科学—岩土工程]