长江科学院院报 ›› 2015, Vol. 32 ›› Issue (9): 146-152.DOI: 10.11988/ckyyb.20140236

• 岩土工程 • 上一篇    下一篇

基于时移小波-灰色理论的边坡位移预测模型研究

郭海庆1a,1b,张敏1a,1b,黄涛1a,1b,艾纯斌2   

  1. 1.河海大学a.岩土工程科学研究所;b.岩土力学与堤坝工程教育部重点实验室,南京 210098; 2.中国市政工程东北设计研究总院有限公司 海南洋浦分公司,海南 洋浦 570125
  • 收稿日期:2014-03-30 出版日期:2015-09-20 发布日期:2015-09-10
  • 作者简介:郭海庆(1974-),男,山东鄄城人,副教授,博士,主要研究复杂岩土体在应力、渗流以及多场耦合作用下的变形与稳定,岩土体的变形监测与数值计算,(电话)13913837880(电子信箱)ghq@hhu.edu.cn。
  • 基金资助:
    国家自然科学基金项目(21372329);中央高校基本科研业务费专项资金资助(2015B06014)

Prediction Model for Rock-Soil Slope Based on Time Shift Wavelet and Grey Theory

GUO Hai-qing1,2, ZHANG Min1,2, HUANG Tao1,2, AI Chun-bin3   

  1. 1.Geotechnical Research Institute, Hohai University, Nanjing 210098, China;
    2.Key Laboratory of Ministry of Education for Geomechanics and Embankment Engineering, Hohai University, Nanjing 210098, China;
    3.Yangpu Hainan Branch, China Northeast Municipal Engineering Design & Research Institute Co., Ltd., Yangpu 570125, China
  • Received:2014-03-30 Online:2015-09-20 Published:2015-09-10

摘要: 边坡的变形稳定性问题是土木工程建设中亟待解决的问题之一。大量研究表明,用实测的边坡位移时间序列预测边坡未来变形更为准确。但外界因素可能使数据产生误差,需去噪处理,才能使监测数据更有使用价值。结合时移小波去噪和灰色理论,对锦屏一级水电站边坡位移监测数据进行研究,提出了时移小波系数相关性去噪及小波-MGM(1,n)预测模型。该模型通过对小波尺度系数和近似系数的分解与重构来模拟真实信号,进而预测边坡的深度位移曲线。经验证预测曲线与实测曲线很接近,为边坡的治理和防护提供了一定的参考依据。

关键词: 边坡稳定性, 小波分析, 去噪处理, 灰色理论, 数据处理

Abstract: Slope's deformation stability has been a pressing issue in civil engineering. A large number of studies have shown that it is accurate to predict slope deformation using measured displacement-time series. But de-noising of data is needed because of errors caused by external factors. In the present paper we put forward a time shift wavelet coefficient correlation de-noising and wavelet-MGM(1, n) model. The model is based on time shift wavelet theory and gray theory. The slope displacement data of Jinping first stage hydropower station is taken as an example. Through decomposition and reconstruction of wavelet scale coefficients and approximation coefficients, the real signal is simulated, and the slope elevation-displacement curve is predicted. Validation proves that the prediction curve is very close to the measured curve.

Key words: slope stability, wavelet analysis, de-noising processing, grey theory, data processing

中图分类号: