Development of Ground Movements Due to a Shield Tunnelling Prediction Model Using Random Forests

In order to predict the exact amount of maximum surface settlement value, this paper presents a method to predict ground movement above tunnels with random forests (RF). Surface settlement above a tunnel due to a tunnel construction is predicted with the help of input variables that have direct physical significance. The RF-based model is developed by free R programs, trained and tested with parameters obtained from the detailed investigation of different tunnel projects published in literature. The maximum settlement is taken as a function of tunnel diameter, depth to the tunnel axis, cohesion, internal friction angle, compressibility modulus of soil, grouting pressure, percent tail voild grout filling, thrust force and advance rate for shield tunneling. A repeated 5-fold cross-validation procedure (10 repeats) is implemented to determine the optimal parameter values during modeling and an external testing set is employed to validate the prediction performance of models. Two performance measures namely R² and RMSE have been employed. The RF demonstrated a promising result and predicted the desired goal fairly successfully.