A Prediction Method of Ground Deformation in Subway Tunnel Construction based on Gaussian Process Regression

Abstract

With the continuous development of urban traffic, the coverage area of the subway has gradually expanded, and the number of subway tunnel construction has increased significantly. To ensure that the construction process does not affect the ground, the ground deformation prediction method of subway tunnel construction is often used to predict the ground conditions. The method is not strong enough to analyze the ground deformation mechanism, which causes the problem that the predicted result is quite different from the measured result. Therefore, a prediction method of ground deformation based on Gaussian process regression is proposed. Set the ground displacement monitoring process, combined with the characteristics of tunnel construction, complete the monitoring and analysis of ground displacement data. The mechanism of ground deformation is analyzed, and combined with the grey prediction model, the ground deformation is predicted. Based on Gaussian process regression equation and covariance calculation results, a ground deformation prediction model is established. So far, the design of ground deformation prediction method for subway tunnel construction based on Gaussian process regression is completed. Constructing the experimental link, the comparison shows that the prediction result of this method is closer to the actual measurement result, which shows that this method is more effective for ground deformation prediction.