基于历史坡面位移监测数据的滑坡变形预测

Landslide Deformation Prediction Based on Historical Slope Displacement Monitoring Data

根据最优加权组合理论,结合多个适应不同曲线特征的数学模型,建立最优加权组合模型用于滑坡变形的预测。基于实际滑坡监测数据分析,对加权组合模型及单一模型在数据拟合精度及预测精度方面进行了分析比较。拟合数据的结果表明,加权组合模型的拟合精度最高,预测值与实际值比较吻合,因此加权组合模型可提高滑坡变形临近时段的预测精度,预测数据可用于滑坡的临滑预警。

According to the optimal weighted combination theory and considering several mathematical models adapted to different curve characteristics,the optimal weighted combination model is established to predict landslide deformation.Based on the analysis of actual landslide monitoring data,the data fitting accuracy and prediction accuracy of single model and combined model are analyzed and compared.According to the residual of fitting data,the fitting accuracy of optimum weighted combined model is the highest in any single model,and the predicted value is in good agreement with the actual value.Therefore,the weighted combination model can improve the prediction accuracy of landslide deformation in the near period,and the prediction data can be used for landslide impending warning.