Multi-objective optimization control for tunnel boring machine performance improvement under uncertainty

The tunnel boring machine (TBM) is an important and common construction method for urban subways, and it requires a detailed and rational control strategy to ensure the safety and efficiency of TBM excavation. Multiple objectives are required for shield tunneling; however, the control of TBM parameters is a complex and difficult problem under frequently encountered unforeseen geological conditions. Hence, a multi-objective optimization framework has been proposed to provide suggested TBM operational parameters for decision making under uncertainty. A Grey Wolf Optimizer-Generalized Regression Neural Network (GWO-GRNN) model has been developed to predict the TBM performance under different TBM operating parameters and geological conditions. Then, the nondominated sorting genetic algorithm (NSGA-II) is introduced to solve the multi-objective optimization problem and obtain the final decision-making solutions. To indicate the applicability of the proposed multi-objective optimization (MOO) framework, the Wuhan San-Yang Road Highway-Rail Tunnel Shield Project was adopted as an example. Results show that the GWO-GRNN model is in good agreement with the experimental measurements to predict the advance speed and ground settlement, with R2 values of 0.97 and 0.91, respectively. Additionally, the results of NSGA-II optimization show that the proposed framework can realize the optimization of multiple objectives under different geological conditions. The results of this research are able to generate the optimal solutions for TBM operators, which can improve decision making when conflicting TBM excavation objectives exist.