Influence analysis and relationship evolution between construction parameters and ground settlements induced by shield tunneling under soil-rock mixed-face conditions

This study investigates the influences of the key shield parameters on tunneling-induced ground settlements considering the dynamic excavation process. The advanced machine learning method in conjunction with principal component analysis is applied to evaluate the nonlinear relationships between ground settlements and the multi-parameters in the soil-rock composite ground. Considering the interaction among multi-factor, the nonlinear relationships of excavation characteristic parameters are developed by the correlation analysis to quantify shield machine performance in the mixed ground. Predictions on ground settlements in the soil-rock mixed-face are conducted in good agreement with the in-site measured data from the Nanjing Metro Line No. S7 tunnel. The implementation effects of the Light-GBM based model are evaluated by RMSE and R2. Subsequently, the influence weights of the most important influencing factors are ranked to control the parameters, which will be beneficial to optimize the excavation scheme during the dynamic construction process. Finally, the empirical formulas are proposed according to the high-correlation parameters as well as the measured datasets by the nonlinear multiple regression process, which can provide guidance for shield tunnel operation in the soilrock composite ground.