Application of BP Neural Network in Prediction of Ground Settlement in Shield Tunneling

Abstract

Ground settlement is one of the control indicators that the shield tunnel construction pays attention to. Ground settlement is affected by multiple construction processes of shield excavation face cutting, shield crossing and shield tail protruding. This paper adopts a data-driven method and based on actual engineering data to establish the relationship between the disturbance stage of the soil in front of the shield cut, the disturbance stage of the soil above the shell, and the disturbance stage of the soil behind the shell after passing through the shell and the final settlement of the ground. The neural network model and sensitivity analysis are used to explore the influence of settlement at different stages on the final settlement. Preliminary summary of the influence degree and law of different geological conditions and the settlement value of each stage on the final consolidation settlement provides a reference for the setting of the settlement value control target value and the design of the control strategy in the actual construction process.