混凝土面板堆石坝裂缝走势预测算法研究OA

A concrete-faced rockfill dam is an important structural type of water conservancy projects.During long-term operation,the dam is prone to cracking due to the combined effects of water pressure,temperature,and other factors,which affect its structural safety and stability.In the prediction of crack propagation trends in rockfill dams,crack trend data are complex time series with significant temporal dependence and spatial correlation.Their spatiotemporal characteristics are difficult to fully exploit,resulting in limited prediction accuracy.In this paper,a hybrid model combining long short-term memory(LSTM)and a convolutional neural network(CNN)is developed to automatically learn and extract features,such as complex spatial correlations,from monitoring data,thus providing reliable input for crack propagation trend prediction.The extracted features are projected into the output space of the convolutional neural network,and the feature information is integrated through a nonlinear transformation to output the predicted values of crack deformation.Based on existing research methods,a comparative experiment is conducted using actual deformation data.The results show that the LSTM-CNN method can accurately locate cracks,with the error between predicted and measured values less than 0.1 mm.The prediction model has high accuracy and can effectively ensure the overall stability of the dam.