Intelligent prediction method for longitudinal deformation of shield tunnel based on physical information neural network

Abstract

The application discloses a shield tunnel longitudinal deformation intelligent prediction method based on a physical information neural network, and the method comprises the following steps: collecting original heterogeneous data, performing data cleaning on the original heterogeneous data, selecting initial characteristic variables from the cleaned data, drawing a correlation heat map according to the initial characteristic variables, and obtaining final characteristic variables according to the correlation heat map; performing standardization processing on the final characteristic variables and dividing the final characteristic variables into a training set and a test set; constructing a physical information neural network model, obtaining a coincidence formula according to a Timoshenko beam and a longitudinal continuous equivalent theory, and constructing a loss function according to the coincidence formula; training and testing the physical information neural network model by using the training set, the loss function and the test set, and obtaining a target neural network model; and performing longitudinal deformation prediction on the shield tunnel by using the target neural network model, and obtaining longitudinal deformation results of the shield tunnel. The application effectively improves the accuracy of intelligent prediction of longitudinal deformation of the shield tunnel.

Images