Bridge Monitoring Data Reconstruction Based On N-BEATS-Transformer Model

In bridge monitoring practice, the loss of sensor data due to uncertainty occurs from time to time, in order to ensure that the data can be used for subsequent bridge analysis, this paper combines the N-BEATS algorithm with the transformer network on the basis of the transformer network, constructs the N-BEATS-Transformer combinatorial model, and uses the deflection data of Meixi River Bridge as the dataset, and conducts training and testing of the model. The experimental results show that the combined N-BEATS-Transformer model has higher prediction accuracy, with lower RMSE and MAE compared to the single Transformer, which are 0.1138 and 0.0824, respectively. The combined model has been used in bridge monitoring systems, and practice has shown that the combined model can meet the needs of bridge deflection data reconstruction, which is conducive to the use of data for subsequent bridge analysis.