基于支持向量机和改进BP神经网络的路基边坡稳定性研究
|
摘要
针对京新高速公路项目在建设中遇到的裂缝、滑移、倾倒等大量边坡稳定性问题,为了探讨边坡岩土体参数与边坡稳定性间的相关关系,以及保证研究项目路段在运营期间的行车安全,实现公路网尤其是山区公路的安全、高效、便捷运行,在已有研究的基础上,分别建立了支持向量机以及附加动量因子mc而改进后的BP神经网络两种边坡稳定性预测模型。通过引入45个训练样本,对5个工程边坡实例的安全系数进行预测计算,分析了两种模型的平均误差和最大误差,比较了两种模型的预测精度和适用范围,并且对京新高速公路胶泥湾至冀晋界路段的工程边坡稳定性进行了预测。结果显示,样本训练阶段,支持向量机和BP神经网络两种模型均具有较高的模拟精度,而BP神经网络更优;在样本预测阶段,支持向量机的预测精度明显优于BP网络;当随着样本容量不断增大时,两种计算模型的预测精度也逐渐提高;通过结果可以得出,支持向量机预测模型有较强的外推能力和预测计算的有效性,可以更好地描述边坡稳定性复杂的非线性关系,更适用于边坡稳定性的预测分析。
Aiming at a large number of slope stability problems such as cracks, slips and dumping encountered in the construction of Beijing-Xinjiang expressway project,in order to explore the correlation between rock mass parameters and slope stability,to ensure the driving safety of the research project sections during operation,and to realize the safe,effective and convenient operation of the highway in mountain area,on the basis of the existing research,2 corresponding slope stability prediction models based on SVM and improved BP neural network with additional momentum factor mc are established respectively. By introducing45 training samples,predictive calculation of the safety factors for 5 engineering slope instances is conducted,the mean errors and maximum errors of the 2 models are analysed,the prediction accuracies and application ranges of the 2 models are compared,and the stability of the slope of Jiaoniwan to Hebei-Shanxi boundary section in Beijing-Xinjiang expressway is predicted. The result shows that( 1) at sample training stage,both the 2 models have higher simulation accuracy,and the BP network is the better;( 2) at sample prediction stage,the prediction accuracy of SVM is obviously better than that of BP network;( 3) as the sample size increases,the prediction accuracies of the 2 computational models increase gradually. From the result it can be drawn that the SVM prediction model has a strong extrapolation ability and the effectiveness of prediction calculation,it can describe the nonlinear relation of slope stability better,which is more suitable for prediction analysis of slope stability.
Aiming at a large number of slope stability problems such as cracks, slips and dumping encountered in the construction of Beijing-Xinjiang expressway project,in order to explore the correlation between rock mass parameters and slope stability,to ensure the driving safety of the research project sections during operation,and to realize the safe,effective and convenient operation of the highway in mountain area,on the basis of the existing research,2 corresponding slope stability prediction models based on SVM and improved BP neural network with additional momentum factor mc are established respectively. By introducing45 training samples,predictive calculation of the safety factors for 5 engineering slope instances is conducted,the mean errors and maximum errors of the 2 models are analysed,the prediction accuracies and application ranges of the 2 models are compared,and the stability of the slope of Jiaoniwan to Hebei-Shanxi boundary section in Beijing-Xinjiang expressway is predicted. The result shows that( 1) at sample training stage,both the 2 models have higher simulation accuracy,and the BP network is the better;( 2) at sample prediction stage,the prediction accuracy of SVM is obviously better than that of BP network;( 3) as the sample size increases,the prediction accuracies of the 2 computational models increase gradually. From the result it can be drawn that the SVM prediction model has a strong extrapolation ability and the effectiveness of prediction calculation,it can describe the nonlinear relation of slope stability better,which is more suitable for prediction analysis of slope stability.
引文
[1]胡厚田.边坡地质灾害的预测预报[M].成都:西南交通大学出版社,2001.HU Hou-tian.Prediction and Forecast of Slope Geological Hazard[M].Chengdu:Southwest Jiaotong University Press,2001.
[2]李宁,姚显春,张承客.岩质边坡动力稳定性分析的几个要点[J].岩石力学与工程学报,2012,31(5):873-881.LI Ning,YAO Xian-chun,ZHANG Cheng-ke.Several Points in Dynamic Stability Analysis of Rock Slope[J].Chinese Journal of Rock Mechanics and Engineering,2012,31(5):873-881.
[3]唐小松,李典庆,曹子君,等.有限数据条件下边坡可靠度分析的Bootstrap方法[J].岩土力学,2016,37(3):893-901,911.TANG Xiao-song,LI Dian-qing,CAO Zi-jun,et al.ABootstrap Method for Analyzing Slope Reliability Based on Limited Shear-strength Parameter Data[J].Rock and Soil Mechanics,2016,37(3):893-901,911.
[4]胡厚田.崩塌落石综合预测方法的研究[J].铁道工程学报,1996,6(2):182-190.HU Hou-tian.Synthetical Prediction of Landfall and Rockfall[J].Journal of Railway Engineering Society,1996,6(2):182-190.
[5]邱向荣,袁仁茂,许伟文.公路边坡灾害危险性预测模糊综合评判法[J].水土保持研究,2003,10(3):26-28,36.QIU Xiang-rong,YUAN Ren-mao,XU Wei-wen.Fuzzy Mathematics Method for Evaluation of Slope Stability[J].Research of Soil and Water Conservation,2003,10(3):26-28,36.
[6]赵洪波,冯夏庭.支持向量机函数拟合在边坡稳定性估计中的应用[J].岩石力学与工程学报,2003,22(2):241-245.ZHAO Hong-bo,FENG Xia-ting.Application of Support Vector Machines Function Fitting in Slope Stability Evaluation[J].Chinese Journal of Rock Mechanics and Engineering,2003,22(2):241-245.
[7]黄发明,殷坤龙,张桂荣,等.多变量PSO-SVM模型预测滑坡地下水位[J].浙江大学学报:工学版,2015,49(6):1193-1200.HUANG Fa-ming,YIN Kun-long,ZHANG Gui-rong,et al.Prediction of Groundwater Level in Landslide Using Multivariable PSO-SVM Model[J].Journal of Zhejiang University:Engineering Science Edition,2015,49(6):1193-1200.
[8]周爱红,王帅伟,袁颖,等.岩质边坡落石运动特征参数分析及SVM预测模型[J].公路交通科技,2017,34(3):20-25.ZHOU Ai-hong,WANG Shuai-wei,YUAN Ying,et al.Analysis on Characteristic Parameters of Rock Slope Rockfall Movement and SVM Prediction Model[J].Journal of Highway and Transportation Research and Development,2017,34(3):20-25.
[9]赵志刚,左仕,李清.基于支持向量机与神经网络法的路基沉降预测对比研究[J].路基工程,2015(4):15-19.ZHAO Zhi-gang,ZUO Shi,LI Qing.Comparative Study on Prediction of Subgrade Settlement Based on Support Vector Machine and Neural Network Methods[J].Subgrade Engineering,2015(4):15-19.
[10]肖大海,谢全敏,杨文东.基于多变量的集成预测模型在隧道拱顶沉降变形预测中的应用[J].公路交通科技,2017,34(12):90-96.XIAO Da-hai,XIE Quan-min,YANG Wen-dong.Application of Integrated Forecasting Model Based on Multivariable in Tunnel Vault Settlement Forecasting[J].Journal of Highway and Transportation Research and Development,2017,34(12):90-96.
[11]于国强,张茂省,王根龙,等.支持向量机和BP神经网络在泥石流平均流速预测模型中的比较与应用[J].水力学报,2012,43(增2):105-110.YU Guo-qiang,ZHANG Mao-sheng,WANG Gen-long,et al.Application and Comparison of Prediction Models of Support Vector Machines and Back-propagation Artificial Neural Network for Debris Flow Average Velocity[J].Journal of Hydraulic Engineering,2012,43(S2):105-110.
[12]NILSSON P,UVO C B,BERNDTSSON R.Monthly Runoff Simulation:Comparing and Combining Conceptual and Neural Network[J].Models Journal of Hydrology,2006,321(1):344-363.
[13]王超.工程高边坡稳定性预测方法研究[D].北京:北京交通大学,2009.WANG Chao.Study on Prediction Methods for High Engineering Slope[D].Beijing:Beijing Jiaotong University,2009.
[14]XUE Xin-hua,YANG Xing-guo,CHEN Xin.Application of a Support Vector Machine for Prediction of Slope Stability[J].Science China:Technological Sciences,2014,57(12):2379-2386.
[15]高大钊.关于岩土设计参数标准值计算公式的讨论[J].工程勘察,1996,24(3):5-8.GAO Da-zhao.Discussion on Calculation Formula of Standard Value of Geotechnical Design Parameters[J].Geotichnical Investigation and Surveying,1996,24(3):5-8.
[16]杨建民,张丹蕾,秦军.考虑渗流作用的土坡稳定分析Fellenius法和简化Bishop法[J].工业建筑,2017,47(12):111-120.YANG Jian-min,ZHANG Dan-lei,QIN Jun.Fellenius Method and Simplified Bishop Method for Slope Stability Analysis Considering Seepage Effect[J].Industrial Construction,2017,47(12):111-120.
[17]唐高朋,赵炼恒,李亮,等.基于MATLAB的边坡稳定性极限上限分析程序开发[J].岩土力学,2013,34(7):2091-2098.TANG Gao-peng,ZHAO Lian-heng,LI Liang,et al.Program Development for Slope Stability Using MATLABSoftware and Upper Bound Limit Analysis[J].Rock and Soil Mechanics,2013,34(7):2091-2098.
[18]ZHANG Chun-hua,TIAN Ying-jie,DENG Nai-yang.The New Interpretation of Support Vector Machines on Statistical Learning Theory[J].Science China Mathematics,2010,53(1):151-164.
[2]李宁,姚显春,张承客.岩质边坡动力稳定性分析的几个要点[J].岩石力学与工程学报,2012,31(5):873-881.LI Ning,YAO Xian-chun,ZHANG Cheng-ke.Several Points in Dynamic Stability Analysis of Rock Slope[J].Chinese Journal of Rock Mechanics and Engineering,2012,31(5):873-881.
[3]唐小松,李典庆,曹子君,等.有限数据条件下边坡可靠度分析的Bootstrap方法[J].岩土力学,2016,37(3):893-901,911.TANG Xiao-song,LI Dian-qing,CAO Zi-jun,et al.ABootstrap Method for Analyzing Slope Reliability Based on Limited Shear-strength Parameter Data[J].Rock and Soil Mechanics,2016,37(3):893-901,911.
[4]胡厚田.崩塌落石综合预测方法的研究[J].铁道工程学报,1996,6(2):182-190.HU Hou-tian.Synthetical Prediction of Landfall and Rockfall[J].Journal of Railway Engineering Society,1996,6(2):182-190.
[5]邱向荣,袁仁茂,许伟文.公路边坡灾害危险性预测模糊综合评判法[J].水土保持研究,2003,10(3):26-28,36.QIU Xiang-rong,YUAN Ren-mao,XU Wei-wen.Fuzzy Mathematics Method for Evaluation of Slope Stability[J].Research of Soil and Water Conservation,2003,10(3):26-28,36.
[6]赵洪波,冯夏庭.支持向量机函数拟合在边坡稳定性估计中的应用[J].岩石力学与工程学报,2003,22(2):241-245.ZHAO Hong-bo,FENG Xia-ting.Application of Support Vector Machines Function Fitting in Slope Stability Evaluation[J].Chinese Journal of Rock Mechanics and Engineering,2003,22(2):241-245.
[7]黄发明,殷坤龙,张桂荣,等.多变量PSO-SVM模型预测滑坡地下水位[J].浙江大学学报:工学版,2015,49(6):1193-1200.HUANG Fa-ming,YIN Kun-long,ZHANG Gui-rong,et al.Prediction of Groundwater Level in Landslide Using Multivariable PSO-SVM Model[J].Journal of Zhejiang University:Engineering Science Edition,2015,49(6):1193-1200.
[8]周爱红,王帅伟,袁颖,等.岩质边坡落石运动特征参数分析及SVM预测模型[J].公路交通科技,2017,34(3):20-25.ZHOU Ai-hong,WANG Shuai-wei,YUAN Ying,et al.Analysis on Characteristic Parameters of Rock Slope Rockfall Movement and SVM Prediction Model[J].Journal of Highway and Transportation Research and Development,2017,34(3):20-25.
[9]赵志刚,左仕,李清.基于支持向量机与神经网络法的路基沉降预测对比研究[J].路基工程,2015(4):15-19.ZHAO Zhi-gang,ZUO Shi,LI Qing.Comparative Study on Prediction of Subgrade Settlement Based on Support Vector Machine and Neural Network Methods[J].Subgrade Engineering,2015(4):15-19.
[10]肖大海,谢全敏,杨文东.基于多变量的集成预测模型在隧道拱顶沉降变形预测中的应用[J].公路交通科技,2017,34(12):90-96.XIAO Da-hai,XIE Quan-min,YANG Wen-dong.Application of Integrated Forecasting Model Based on Multivariable in Tunnel Vault Settlement Forecasting[J].Journal of Highway and Transportation Research and Development,2017,34(12):90-96.
[11]于国强,张茂省,王根龙,等.支持向量机和BP神经网络在泥石流平均流速预测模型中的比较与应用[J].水力学报,2012,43(增2):105-110.YU Guo-qiang,ZHANG Mao-sheng,WANG Gen-long,et al.Application and Comparison of Prediction Models of Support Vector Machines and Back-propagation Artificial Neural Network for Debris Flow Average Velocity[J].Journal of Hydraulic Engineering,2012,43(S2):105-110.
[12]NILSSON P,UVO C B,BERNDTSSON R.Monthly Runoff Simulation:Comparing and Combining Conceptual and Neural Network[J].Models Journal of Hydrology,2006,321(1):344-363.
[13]王超.工程高边坡稳定性预测方法研究[D].北京:北京交通大学,2009.WANG Chao.Study on Prediction Methods for High Engineering Slope[D].Beijing:Beijing Jiaotong University,2009.
[14]XUE Xin-hua,YANG Xing-guo,CHEN Xin.Application of a Support Vector Machine for Prediction of Slope Stability[J].Science China:Technological Sciences,2014,57(12):2379-2386.
[15]高大钊.关于岩土设计参数标准值计算公式的讨论[J].工程勘察,1996,24(3):5-8.GAO Da-zhao.Discussion on Calculation Formula of Standard Value of Geotechnical Design Parameters[J].Geotichnical Investigation and Surveying,1996,24(3):5-8.
[16]杨建民,张丹蕾,秦军.考虑渗流作用的土坡稳定分析Fellenius法和简化Bishop法[J].工业建筑,2017,47(12):111-120.YANG Jian-min,ZHANG Dan-lei,QIN Jun.Fellenius Method and Simplified Bishop Method for Slope Stability Analysis Considering Seepage Effect[J].Industrial Construction,2017,47(12):111-120.
[17]唐高朋,赵炼恒,李亮,等.基于MATLAB的边坡稳定性极限上限分析程序开发[J].岩土力学,2013,34(7):2091-2098.TANG Gao-peng,ZHAO Lian-heng,LI Liang,et al.Program Development for Slope Stability Using MATLABSoftware and Upper Bound Limit Analysis[J].Rock and Soil Mechanics,2013,34(7):2091-2098.
[18]ZHANG Chun-hua,TIAN Ying-jie,DENG Nai-yang.The New Interpretation of Support Vector Machines on Statistical Learning Theory[J].Science China Mathematics,2010,53(1):151-164.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |