基于时间序列分析法的滑坡变形特征研究

Study of Landslide Deformation Characteristics Based on Time Series Analysis

应用时间序列分析方法建立滑坡位移ARIMA预报模型。采用差分平稳,自回归AR模型和移动平均MA模型对滑坡位移进行预测,得到了该滑坡监测点TP1的预报模型为ARIMA(2,2,1),然后分析对比实测与预测位移–时间曲线之间的关系。计算结果能够较好地体现出滑坡在外界诱发因素作用下位移的发展变化趋势,说明所建滑坡位移预测预报模型效果较好,在滑坡位移预测中是有效可行的。

An ARIMA prediction model of landslide displacement based on the landslide displacement monitoring data in Baishi landslide, Beichuan County, is established by using time series analysis method. Based on the self regression AR model and moving average MA model of landslide displacement prediction, the difference stationary process is used for calculating the partial displacement which, in combination with the forecasting values, obtains the total displacement prediction value and forecast model ARIMA （2,2,1） of the landslide monitoring point TP1. A displacement - time curve is obtained from measured and predicted values. The calculation results can well reflect the trend of the development and change of landslide displacement under the action of external factors which indicates that this model is effective and feasible in the prediction of landslide displacement.