基于改进的支持向量机隧道变形预测模型

摘要
图/表
参考文献
相关文章 (5)

全文: PDF (857 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要研究目的：用于消除地铁隧道变形监测数据受环境等因素引起的噪声影响，为提高变形数据的外推预测能力，提出基于离散小波消噪和动态在线滑移窗的支持向量机预测方法。研究结论：（1）利用离散小波变换对地铁隧道变形数据进行消噪处理，提取低频有效工程信息；（2）引入动态滑移窗技术，通过二阶双重滑移窗口的动态调整、更新，确定隧道变形预测模型的最终参训样本和实训样本数据，提高变形数据的有效利用率；（3）利用上海地铁隧道变形数据，从稳态与非稳态两个工况下分别对预测模型进行验证，检测模型的预测效果，结果表明该方法具有消噪效果好、预测精度高等特点，同时该预测模型可转为工程化应用，对隧道预测模型的搭建具有指导意义。

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章

关键词 ：隧道变形, 支持向量机, 小波分析, 滑移窗口, 预测

Abstract：Abstract：Research purposes： In order to remove the noise of metro tunnel deformation monitoring data which affected by environmental factors and increase the extrapolated predictive ability of deformation data, a new prediction method based on discrete wavelets and a modified support vector machine (SVM) is proposed.

Research conclusions：(1) The noise of tunnel deformation data can be removed by discrete wavelet analysis and the low frequency effective information can be acquired. (2) A dynamic on-line sliding window technique is introduced. Additionally, the final training samples and the training samples of the tunnel deformation prediction model could be determined by utilizing the two orders of sliding windows which aiming at dynamically adjusting and updating the number of samples. This could improve data utilization rate. (3) The performance of prediction model has been validated with Shanghai metro deformation data under steady state and unsteady state. The results show that this method has good de-noising effect and high precision. Also, the prediction model could be transferred to engineering application and may give references to the construction of tunnel prediction model.

Key words： Key words: tunnel deformation support vector machine wavelet analysis sliding window prediction

基金资助:国家自然科学基金项目（51375345）；上海科学技术委员会资助项目（08201202103）

作者简介: 周奇才， 1962年出生，男，教授，博士生导师。

引用本文:

周奇才范思遐赵炯熊肖磊. 基于改进的支持向量机隧道变形预测模型[J]. 铁道工程学报, 2015, 32(3): 67-.
ZHOU Qi-cai,FAN Si-xia, ZHAO Jiong, XIONG Xiao-lei. Tunnel Deformation Prediction Based on Modified Support Vector Machine. 铁道工程学报, 2015, 32(3): 67-.

链接本文:

http://tdgcxb.crec.cn/CN/ 或 http://tdgcxb.crec.cn/CN/Y2015/V32/I3/67

［1］冯夏庭. 智能岩石力学导论［M］. 北京:科学出版社，2000.
Feng Xiating. Introduce of Intelligent Rock Mechanics［M］. Beijing: Science Press, 2000.
［2］李广信．高等土力学［M］．北京：清华大学出版社, 2004.
Li Guangxin. Advanced Soil Mechanics［M］. Beijing: Tsinghua University Press, 2004.
［3］韦凯. 基于蚁群算法的软土盾构隧道长期不均匀沉降预测模型［D］. 上海：同济大学，2009.
Wei Kai. Forecast of Uneven Long-term Settlement of Metro Tunnel on the Basis of Ant Colony Optimization in Soft Soil［D］. Shanghai：Tongji University,2009.
［4］Cortes C, Vapnik V. Support Vector Networks［J］. Machine Learning, 1995(3):273-297.
［5］Ding L Y, Ma L, Luo H B, etc. Wavelet Analysis for Tunneling Induced Ground Settlement Based on a Stochastic Model［J］. Tunnelling and Underground Space Technology,2011(5):619-628.
［6］Hui X, Bi W H, Hou X L, et.al. Reliability Analysis Using Radial Basis Function Networks and Support Vector Machines［J］. Computers and Geotechnics. 2011(2):178-186.
［7］Stefano M, Giuseppe J. Terminated Ramp-support Vector Machines: a Nonparametric Data Dependent Kernel［J］. Neural Networks, 2006(10):1597-1611．
［8］Liu H, Wen S, Li W, Xu C, Hu C. Study on Identification of Oil/gas and Water Zones in Geological Logging Base on Support-vector Machine［J］. Fuzzy Inform Eng,2009(2):849-857.
［9］范思遐,周奇才,熊肖磊,等. 基于粒子群与支持向量机的隧道变形预测模型［J］. 计算机工程与应用，2014(5):6-10.
Fan Sixia, Zhou Qicai, Xiong Xiaolei, etc. A Tunnel Deformation Prediction Model Based on Support Vector Machine with Particle Swarm Optimization Algorithm［J］. Computer Engineering and Application,2014(5):6-10.
［10］周奇才,范思遐, 赵炯, 等. 基于Multi-agent的隧道变形监测系统［J］. 铁道工程学报，2014(9):77-82.
Zhou Qicai, Fan Sixia, Zhao Jiong,etc. Tunnel Deformation Monitoring System Based on Multi-agent Technique［J］. Journal of Railway Engineering Society, 2014(9):77-82.