A tunneling speed enhancement method for super-large-diameter shield machines considering strata heterogeneity

Abstract

The geological environments faced by super-large-diameter shields are more complex than those encountered by regular-diameter shields. Incorrectly set excavation parameters can lead to increased construction costs and even serious engineering accidents. To ensure safe and efficient excavation processes for shield machines in complex strata, in this paper, which is based on the super-large-diameter shield project of the Jiangyin–Jingjiang Yangtze River Tunnel, a method for proportionally restoring the heterogeneous characteristics of composite strata, which is named the “tunnel face color image” method, is proposed for the first time. Utilizing machine learning and the grey wolf optimizer, models for predicting the tunneling speed and constrained items of the shield machine are established. On this basis, an improved grey wolf optimizer is further developed to construct an adaptive decision-making system for setting the main control parameters of the shield with the objective of maximizing the tunneling speed while satisfying the constraints imposed on the cutterhead torque and attitude deviations. The results show that the tunnel face color image method can effectively extract geological information from each shield cycle and use it as an input for the prediction model, resulting in an average absolute error of 0.916 mm/min for the most important tunneling speed prediction result and a determination coefficient of 0.879, thus outperforming other geological parameter processing methods. The adaptive decision-making system for setting the main control parameters of the shield, which is based on the improved grey wolf optimizer, is capable of accurately solving for the optimal operating parameters for each shield cycle with an optimization time that is shorter than those of the particle swarm optimization algorithm, genetic algorithm, and artificial fish swarm algorithm. Moreover, according to the optimal operating parameters obtained, the average tunneling speed in each ring of the shield can be increased by 39.9 % while reducing the fluctuation range of the cutterhead torque and making the attitude deviations more convergent.