基于影响因子的GM(1,1)-BP模型在八字门滑坡变形预测中的应用

滑坡的预测预报是世界性难题,特别是如何把外界影响因素定量计入到滑坡中长期预报模型中。在滑坡中长期预报模型适用性分析的基础上,提出了基于影响因子的GM(1,1)-BP模型,该模型能考虑库水位、降雨等外界影响因素。以三峡库区八字门滑坡为例,在八字门滑坡工程地质条件的基础上,通过对位移、库水位、降雨等监测资料的分析,系统研究了八字门滑坡的变形特征、影响因素和变形破坏模式。在确定八字门滑坡影响因子的基础上,分别采用本文模型、GM(1,1)模型和不考虑影响因子的BP神经网络对滑坡位移进行了预测研究,发现本文模型预测精度较高。

Application of GM (1, 1)-BP Model to the Deformation Prediction of Bazimen Landslide in Consideration of Influencing Factors

Landslide prediction is a worldwide problem, especially how to quantitatively consider external influencing factors in the long-term prediction. On the basis of analyzing the applicability of long-term landslide prediction, a GM(1,1) BP model based on influencing factors was presented in which external influencing factors such as rainfall and reservoir water level fluctuation were taken into consideration. Bazimen landslide in the Three Gorges Reservoir was taken as a case study. The deformation characteristics, influencing factors, and failure mode of the landslide were investigated by analyzing the monitoring data of displacement, reservoir water level and rainfall. The influencing factors of Bazimen landslide were firstly determined, and in subsequence, the model established in the present study, the GM(1,1) model, and the BP model which does not include the influencing factors were respectively taken to predict the landslide deformation. Results show that the model established in this research has higher precision.