Research on Pressure Prediction of Hydraulic Climbing Formwork Based on LSTNet
Hydraulic climbing formwork is a type of equipment employed in building construction,predicting its pressure is benefi-cial for monitoring the operational status and providing fault warnings during the bridge construction process.In order to obtain more ac-curate pressure prediction results,a hydraulic climbing formwork pressure prediction model based on the long and short-term time-se-ries network was proposed.The Spearman correlation coefficient method was used to filter out the data strongly correlated with the pres-sure data of the hydraulic climbing formwork for reducing the interference of irrelevant data.The LSTNet model was employed to identify the long-term and short-term dependencies in the pressure data of hydraulic climbing formwork equipment.A linear autoregressive layer was introduced and combined with the nonlinear part of the neural network to improve the prediction accuracy of the network model.Fi-nally,the model was trained using the pressure data collected from the Changtai Yangtze River Bridge Hydraulic Climbing Formwork Project and compared with the LSTM model,LSTM-Attention model and CNN-BiLSTM-Attention model.The results show that in the pressure prediction experiments of hydraulic climbing formwork,the LSTNet model demonstrates good fitting and predictive perform-ance,achieving higher accuracy compared to the other three models.Additionally,the LSTNet model combines linear and nonlinear fea-ture extraction capabilities,which enhances the modeling flexibility and accuracy of time series data and significantly improves predic-tive performance of the model.
deep learninghydraulic climbing formworkonline monitoring platformlong and short-term time-series network(LSTNet)pressure prediction
NETL
NSTL
万方数据
