Prediction for the Settlement of Concrete Face Rockfill Dams Using Optimized LSTM Model via Correlated Monitoring Data

Settlement prediction is of great importance for safety control concrete-face rockfill dams (CFRDs) during the operation stage. However, accuracy achieved by commonly used hydrostatic–seasonal–time (HST) methods, without consideration previous conditions influencing factors, not competitive. Moreover, in most settlement data at each monitoring point are modeled individually; correlation relationships between settlements neglected. In this paper, a method based on an optimized long short-term memory (LSTM) model proposed to predict CFRDs, modeling multiple series with strong simultaneously. method, classified into several categories, firstly according global relevance measure. Then, cuckoo search (CS) algorithm applied optimize hyper-parameters neural network structure LSTM. Ultimately, LSTM utilized same category. Results indicate that has better performance compared model, back propagation (BPNN) and HST single point.