Prediction of maximum ground surface settlement induced by shield tunneling using XGBoost algorithm with golden-sine seagull optimization

In order to avoid damage during shield tunnel construction to adjacent buildings on the ground, it is of paramount importance to precisely predict the Maximum Surface Settlement (MSS) deformation induced by shield tunneling under the surface. In this paper, a hybrid algorithm model incorporating the Extreme Gradient Boosting (XGBoost) with the Golden-sine Seagull Optimization Algorithm (GSOA) is proposed, namely the GSOA-XGBoost model. A total of 323 datasets including torque, penetration rate, thrust, cutterhead rotation speed, slurry pressure, grouting pressure, cover depth, and distance between cutting face and monitored section are taken as input parameters, and the corresponding measured MSS deformation is adopted as the output parameter. This dataset is employed to train the MSS prediction algorithm model. The analysis results indicate that the proposed GSOA-XGBoost hybrid algorithm model is much superior to other models in terms of accuracy, stability and computation time, and strong correlations are discovered between the predicted and measured settlements. Our model can predict the MSS deformation caused by shield tunnel construction with higher reliability, strong robustness and fault-tolerant capacity, and greatly reduces the engineering risk. Simultaneously, the consistency of the slurry pressure and grouting pressure values is calculated by the SOA-XGBoost and the grid search method, which demonstrates the feasibility and rationality of proposed hybrid algorithm model.