滑坡位移非线性时间序列预测模型研究

Landslide Displacement Nonlinear Time Series Prediction Model

从分析滑坡位移的实际变化情况和滑坡位移演变规律的角度出发,将滑坡位移分解为受其自身地质结构属性控制的趋势项位移分量和由外界环境因素影响的周期项位移分量。采用经验模态分解法对滑坡位移趋势项和周期项进行非线性时间序列的分解;在此基础上采用标准GM(1,1)灰色模型、滚动GM(1,1)灰色模型和灰色马尔科夫模型分别对滑坡位移的趋势项和周期项进行预测,将预测结果进行叠加运算,即可得到滑坡位移的预测值。以三峡库区白水河滑坡的位移变化情况为例,通过分析对比滑坡位移的实测值与预测值之间的位移-时间关系曲线,可以很好地预测出滑坡位移的发展变化趋势。这说明对滑坡位移进行时间序列分解,有助于提高滑坡位移的预测精度,并可有效应用于滑坡位移预测的工程实例中。

Based on the changes of landslide displacement and displacement evolution of the landslide,displacement can be divided into trend term displacement controlled by its own geological conditions and periodic term displacement affected by external factors.Empirical mode decomposition is used to separate the trend term displacement and periodic term displacement of the landslide.Based on this,GM（1,1）model,roll GM（1,1）model and gray Markov model are respectively used to forecast the displacement of trend term and periodic term.Total displacement is obtained by adding the calculated predictive displacement value of each sub-stack.Taking Baishuihe landslide in the Three Gorges Reservoir area for example,the comparative analysis of the measured and predicted displacement-time curves shows that the model could better reflect the development trend of landslide.The results demonstrate that the established landslide displacement prediction model is feasible and effective in the landslide displacement prediction.Landslide displacement time series decomposition will help improve the prediction accuracy of landslide displacement.Projects can be effectively applied to prediction of landslide displacement.