关键词: 滑坡位移预测; 时间序列; 灰色Verhulst模型; Gated Recurrent Unit; 八字门滑坡;

Abstract: In recent years,deep learning,long and short time memory networks(LSTM)are often used to monitor landslide displacement.Gated Recurrent Unit(GRU)is an improvement for LSTM.A new neural network model for predicting the landslide displacement was proposed based on time series theory and GRU in this paper.In this model,the moving average method was applied to decompose the cumulative displacement into the trend term and periodic term.The trend displacement was predicted by grey Verhulst model.Considering the influence of rainfall and reservoir water level on landslide displacement,a three-layer Gated Recurrent Unit network and Dense network using Python language were constructed to predict the periodic term.The displacement monitoring data of ZG111 in the Bazimen landslide was used to verify the effectiveness of this method.The results showed that the new method can consider historical information more effectively than GRNN,and the prediction effect was significantly improved,which can provide important reference for landslide prediction.

Key words: landslide displacement prediction; time series; grey Verhulst model; Gated Recurrent Unit; Bazimen landslide;