混凝土面板堆石坝, CNN, LSTM, 变形预测

Aiming at the problems of weak model generalization ability and difficult data acquisition in deformation prediction of concrete face rockfill dams, CNN-LSTM model that integrates convolution and deep learning techniques is proposed. Integrating the advantages of CNN and LSTM, a hybrid prediction model is constructed, and its application effect in dam deformation prediction is explored. To verify the superiority of the model, LSTM and CNN models are also established. The comparison results show that the CNN-LSTM model outperforms the other two models in terms of prediction accuracy, that this method can provide an effective means for dam deformation monitoring.

concrete face rockfill dam, CNN, LSTM, deformation prediction