|
|
- 2018
基于高斯过程回归的变形智能预测模型及应用
|
Abstract:
岩体或建构筑物的变形通常具有复杂性和非线性等特性，一般的回归模型难以精确地进行回归预测，应用高斯过程回归理论对变形监测数据呈现出的非线性特征进行时间序列分析。考虑到监测数据的不断更新和累积，以及超参数与样本集的适应性，首先研究了“递进-截尾式”超参数自动更新模式和训练样本集的选择方法；在此基础上构建了以时间作为输入项的高斯过程回归变形智能预测模型（GPR-TIPM）；将该模型应用于矿山边坡监测点非线性时间序列分析中，通过分析变形趋势，最终采用Matérn 32和平方指数协方差函数相加的方式进行核函数组合。实验结果表明，采用组合核函数的预测性能较单一核函数有所改善，该方法提高了模型的泛化能力，GPR-TIPM模型在短期内的预测效果较理想
| [1] | Stojanovic B, Milivojevic M, Ivanovic M. Adaptive System for Dam Behavior Modeling Based on linear Regression and Genetic Algorithms[J]. <em>Advances in Engineering Software, </em>2013, 65:182-190 |
| [2] | Suwansawat S, Einstein H H. Artificial Neural Networks for Predicting the Maximum Surface Settlement Caused by EPB Shield Tunneling[J]. <em>Tunnelling and Underground Space Technology, </em>2006, 21:133-150 |
| [3] | Zhang H, Wang Y, Li Y. SVM Model for Estimating the Parameters of the Probability-integral Method of Predicting Mining Subsidence[J]. <em>Mining Science and Technology (China), China University of Mining and Technology, </em>2009, 19(3):385-388 |
| [4] | Li S, Zhao H, Ru Z. Deformation Prediction of Tunnel Surrounding Rock Mass Using CPSO-SVM Model[J]. <em>Journal of Central South University, </em>2012, 19(11):3311-3319 |
| [5] | Xia Zineng. Forecasting Method of Nonlinear Time Series of Slopedeeormat Using Gaussian Process[D]. Nanning:Guangxi University,2013(夏自能. 边坡位移非线性时间序列的髙斯过程预测方法[D]. 南宁:广西大学, 2012) |
| [6] | Rasmussen C E, Williams C K. Gaussian Processes for Machine Learning[M]. Cambridge, MA:MIT Press, 2006 |
| [7] | Wu Xueling, Ren Fu, Niu Ruiqing. Spatial Intelligent Prediction of Landslide Hazard Based on Multi-source Data in Three Corges Reservoir Area[J]. <em>Geomatics and Information Science of Wuhan University,</em>2013,38(8):963-968(武雪玲,任福,牛瑞卿.多源数据支持下的三峡库区滑坡灾害空间智能预测[J].武汉大学学报·信息科学版,2013,38(8):963-968) |
| [8] | Li L, Huang G. Procedia Engineering Prediction of Goaf Settlement with Time Sequence of Wavelet Neural Network[J]. <em>Procedia Engineering,</em> 2011, 15:4723-4727 |
| [9] | Lu Jun,Dai Wujiao,Zhang Zhetao.Modeling Dam Deformatin Using Varying Goefficient Regression[J]. <em>Geomatics and Information Science of Wuhan University,</em>2015,40(1):139-142(卢骏,戴吾蛟,章浙涛.大坝变形变系数回归建模[J]. 武汉大学学报·信息科学版, 2015, 40(1):139-142) |
| [10] | Wang Q, Wang C, Xie R. An Improved SCGM(1,m) Model for Multi-point Deformation Analysis[J]. <em>Geosciences Journal, </em>2014, 18(4):477-484 |
| [11] | Huang Shengxiang,Yin Hui,Jiang Zheng.Deformation Monitoring Data Processing[M]. Wuhan:Wuhan University Press,2010(黄声享, 尹晖, 蒋征. 变形监测数据处理[M]. 武汉:武汉大学出版社, 2012) |
| [12] | Su G S. Gaussian Process-based Dynamic Response Surface Method for Estimating Slope Failure Probability[J]. <em>Yantu Lixue/Rock and Soil Mechanics, </em>2014, 35(12):3592-3601 |
| [13] | Zhang Zhenglv,Huang Quanyi, Wen Hongyan. Deformation Monitoring Analysis and Prediction for Engineering Constructions[M]. Beijing:Surveying and Mapping Press, 2007(张正禄, 黄全义, 文鸿雁. 工程的变形监测分析与预报[M]. 北京:测绘出版社, 2007) |
| [14] | Zhang Yan. Gaussina Process Model for Forecasting Andidentifying Nonlinear Behaviorofunderground Engineering Rockmass Anddynamic Intelligent Feedback Analysis[D]. Nanning:Guangxi University,2013(张研. 地下工程岩体非线性行为预测识别的高斯过程模型与动态智能反馈分析[D]. 南宁:广西大学, 2013) |
| [15] | Grelle G, Guadagno I F M. Regression Analysis for Seismic Slope Instability Based on a Double Phase Viscoplastic Sliding Model of the Rigid Block[J]. <em>Landslides, </em>2013, 10(5):583-597 |
| [16] | Chen Xiaopeng, Zhang Qiangyoun, Liu Dawen. Deformation Statistical Regressiong Analysis Model of Slope and Its Application[J]. <em>Chinese Journal of Rock Mechanics and Engineering, </em>2008, 27(S2):3673-3678(陈晓鹏, 张强勇, 刘大文. 边坡变形统计回归分析模型及应用[J]. 岩石力学与工程学报, 2008, 27(S2):3673-3678) |
| [17] | Li P, Tan Z, Yan L, et al. Time Series Prediction of Mining Subsidence Based on a SVM[J]. <em>Mining Science and Technology, China University of Mining & Technology,</em>2011, 21(4):557-562 |
| [18] | Li Dejiang,Hua Xianghong,Li Tao.Researchon Predictive Modeling of Bulding Settlement Based on Suport Vector Mechine[J]. <em>Engineering of Surveying and Mapping, </em>2009, 18(3):29-31(李德江, 花向红, 李涛. 基于支持向量机的建筑物沉降预测模型研究[J]. 测绘工程, 2009, 18(3):29-31) |
| [19] | Pan Ping.Slope Displacement Forecase Based Onwavelet Neural Network[J]. <em>Journal of Chengdu University of Technology:Sci & Technolgy, </em>2006, 33(2):176-180(潘平. 基于小波神经网络理论的边坡位移预测[J]. 成都理工大学学报(自然科学版), 2006, 33(2):176-180) |
| [20] | He Zhikun, Liu Guangbin, Zhao Xijing. Overview of Gaussian Process Regression[J]. <em>Control and Decision, </em>2013, 28(8):1121-1129(何志昆, 刘光斌, 赵曦晶. 高斯过程回归方法综述[J]. 控制与决策, 2013, 28(8):1121-1129) |
| [21] | Su Guoshao, Xiao Yilong. Gaussian Process Method for Slope Reliability Analysis[J]. <em>Chinese Journal of Geotechnical Engineering,</em> 2011, 33(6):916-920(苏国韶, 肖义龙. 边坡可靠度分析的高斯过程方法[J]. 岩土工程学报, 2011, 33(6):916-920) |
