|本期目录/Table of Contents|

[1]何新成,杨山岗.基于LIBSVM算法在隧道围岩分级上的应用[J].公路与自然,2019,26(02):24-29.
 HE Xin cheng,YANG Shan gang.Classification of Tunnel Surrounding Rock Based on LIBSVM Method[J].Highway & Nature,2019,26(02):24-29.
点击复制

基于LIBSVM算法在隧道围岩分级上的应用()
分享到:

《公路与自然》[ISSN:/CN:]

卷:
第26卷
期数:
2019年02
页码:
24-29
栏目:
桥隧工程
出版日期:
2019-05-07

文章信息/Info

Title:
Classification of Tunnel Surrounding Rock Based on LIBSVM Method
作者:
何新成杨山岗
西安方舟工程咨询有限责任公司,陕西 西安 710065
Author(s):
HE XinchengYANG Shangang
Xi’an Fangzhou Engineering Consulting Co., Ltd.Xi’an 710065,China
关键词:
LIBSVM支持向量机隧道围岩分级神经网络
Keywords:
LIBSVMsupport vector machinetunnel surrounding rock classificationneural network
分类号:
U452.1
DOI:
-
文献标志码:
A
摘要:
隧道围岩分级是选择隧道位置、支护类型的依据及指导安全施工关键指标。针对公路隧道围岩分级等方法存在人为因素和不确定因素,基于LIBSVM支持向量机,选取岩石质量指标、完整性指标、饱和单轴抗压强度、纵波波速、弹性抗力系数和结构面摩擦因数6个指标作为特征输入因素量,构建隧道围岩分级非线性映射模型,通过交叉验证设计LIBSVM寻找最优参数的方法对现场勘测的30组隧道围岩数据样本进行学习,7组隧道围岩数据样本进行预测。同时对比了模式识别、BP神经网络、Bayes、Fisher、SVM及GA-SVM分类算法。分析和对比结果表明:采用LIBSVM方法对隧道围岩分级预测准确率达100%,其理论完善、计算简便和泛化能力均优于其他神经网络等算法,说明LIBSVM方法可以较好地用于确定隧道围岩分级。
Abstract:
Classification of tunnel surrounding rock is the basis for selecting tunnel location, supporting type and the key index for guiding safe construction. Aiming at the artificial factors and uncertainties in surrounding rock classification of highway tunnels, based on LIBSVM support vector machine, the rock quality index, integrity index, saturated uniaxial compressive strength, Pwave velocity, elastic drag resistance coefficient and structural surtace friction coefficient are selected as characteristic input factors to construct a non-linear mapping model for surrounding rock classification of highway tunnels, which is verified by cross-validation. The method of LIBSVM is designed to find the optimal parameters. Thirty groups of tunnel surrounding rock data samples are studied and seven groups of tunnel surrounding rock data samples are predicted. At the same time, the classification algorithms of pattern recognition, BP neural network, Bayes, Fisher, SVM and GA-SVM are compared. The analysis and comparison results show that the accuracy of classification and prediction of tunnel surrounding rock by LIBSVM method is 100%, and its theoretical perfection, simplicity of calculation and generalization ability are superior to other neural network algorithms. It shows that LIBSVM method can be better used to determine the classification of tunnel surrounding rock.

参考文献/References:

[1]. BARTON N, LIEN R, LUNDE J. Engineering Classification of Rock Masses for the Design of Tunnel Support[J].Rock Mechanics,1974,6 (4): 189-236. [2]. BIENIAWSKI Z T. Engineering Rock Mass Classifications: A Complete Manual for Engineers and Geologists in Mining,Civil and Petroleum Engineering [J].Petroleum,1976,251 (3) : 357-365. [3]. 吴相金,龚建平.公路隧道围岩分类模糊综合评判[J].公路交通科技,2007,24 (1):118-125. [4]. 段林娣,宋成辉.应用BP 神经网络进行隧道围岩快速分级[J].中国安全科学学报,2010,20(2) :41-45. [5]. 王迎超,孙红月,尚岳全,等. 基于特尔菲—理想点法的隧道围岩分类研究[J]. 岩土工程学报,2010,32 (4):651-656. [6]. 余伟健,高谦,韩阳,等.非线性耦合围岩分类技术及其在金川矿区的应用[J]. 岩土工程学报,2008,30 (5) : 663-669. [7]. 李春萍,郝会兵.煤巷围岩分类的Bayes 判别分析法[J].煤炭学报,2011,36(2) : 304-307. [8]. 姜春露,姜振泉,孙强.基于因子分析与Fisher 判别分析法的隧洞围岩分类研究[J].公路交通科技,2015,32(7) : 98-104. [9]. 宫凤强,李夕兵.距离判别分析法在岩体质量等级分类中的应用[J]. 岩石力学与工程学报,2007,26( 1) : 190-194. [10]. 何浩祥,闫维明,彭凌云. 基于支持向量机的钢筋混凝土桥梁损伤识别.[J].公路交通科技,2008,25(3).65-69. [11]. 袁前飞,蔡从中,肖汉光等.基于人体血液/微量元素含量的SVM癌症辅助诊断.[J].生物医学工程学杂志,2007,24(3):513-518. [12]. GB/T 50218-2014,工程岩体分级标准[S]. [13]. 周翠英,张亮,黄显艺.基于改进BP 网络算法的隧洞围岩分类[J].地球科学—中国地质大学学报,2005,30 (4):480-486. [14]. JTG D70/2-2014,公路隧道设计规范[S].(JTG D70/2-2014,Code For Design of Road Tunnel[S].) [15]. 温廷新,于凤娥,邵良杉等.基于GA-SVM 的隧道围岩分类研究[J].公路交通科技,2018,35 (9) : 63-69.

备注/Memo

备注/Memo:
收稿日期:2019.01.07 作者简介:何新成(1985--),男,陕西西安人,2012年毕业于西安建筑科技大学桥隧专业,硕士,工程师,主要从事公路隧道设计与咨询工作.
更新日期/Last Update: 2019-04-25