滑坡预测中灰色预测模型分析

Analysis of grey prediction model in landslide prediction

由于滑坡预测中灰色预测算法复杂多样,从滑坡引起传感器位移的变化出发,采用灰度相关算法进行滑坡下一阶段位移进行预测分析。以滑坡中各个监测点传感器的位移作为输入数据,分别采用灰色GM(1,1)、GM(2,1)和灰色Verhulst建立模型,根据各个算法的预测结果分析。结果表明,根据建立的3种预测模型,结合平溪特大桥滑坡体位移数据,对其下一阶段滑坡位移进行预测。在总体预测过程中,GM(1,1)、GM(2,1)和灰色Verhulst模型具有良好的拟合性,但在模型中GM(1,1)和GM(2,1)曲线拟合较好,其中GM(2,1)在总体上略优于GM(1,1)。

Because the gray prediction algorithm in landslide prediction is complex and diverse, from the change of sensor displacement caused by landslide, the gray-scale correlation algorithm is used to predict the displacement of the next stage of landslide. Taking the displacement of each monitoring point sensor in the landslide as the input data, the models are established by gray GM(1,1), GM(2,1) and gray Verhulst, respectively, and analyzed according to the prediction results of each algorithm. The results show that according to the established three prediction models, combined with the displacement data of the landslide body of Pingxi Bridge, the displacement of the next stage landslide is predicted. In the overall prediction process, the GM(1,1), GM(2,1) and gray Verhulst models have good fit, but the GM(1,1) and GM(2,1) curve fits in the model. Preferably, GM(2,1) is generally slightly better than GM(1,1).