边坡变形的多因素时变预测模型

在深入分析边坡变形的影响因素及因子模式的基础上，通过融合回归分析和递推模型的优点，结合边坡位移时间序列实测资料，建立了边坡变形的多因素回归时变预测模型，并考虑了降雨和温度变化的影响。通过模型的预测值与实测数据的对比分析，可以监控边坡的变形和稳定性。工程实例表明，该方法具有预测精度较高且实时性强等特点，有较大的应用价值。

TIME SERIES EVOLUTION FORECASTING MODEL OF SLOPE DEFORMATION BASED ON MULTIPLE FACTORS

A lot of factors such as topography, geological conditions and rainfall influence slope deformation. The importance of rainfall is difficult to determine. Based on the analysis of the multiple factors of slope deformation and observation of monitoring deformations, a statistic method to correlate slope deformation and rainfall was developed including the hysteresis of rainfall. Whereafter, by applying observation to ordinary statistic model comprising rainfall, temperature and time factor, the main factors affecting slope deformation were extracted by stepwise regression. According to the factors extracted and time series observation, a recursion model was obtained with multilayer recursion analysis method of dynamic system forecasting. At the end, a time series evolution forecasting model of slope deformation was presented combining with regression model and recursion model. The model syncretizes the merits of regression analysis and recursive analysis for enhancing the capability of forecasting evolution of slope deformation. Through comparison between the forecasting and the observation of monitoring deformations, the deformation and stability of slope can be monitored. The engineering case analysis indicates that the model is effective and has big value for forecasting of slope deformation.