Research on Prediction of EPB Shield Tunneling Parameters Based on LGBM

At present, the determination of tunnel parameters mainly rely on engineering experience and human judgment, which leads to the subjective decision of parameters and an increased construction risk. Machine learning algorithms could provide an objective theoretical basis for tunnel parameter decision making. However, due to the limitations of a machine learning model’s performance and parameter selection methods, the prediction model had poor prediction results and low reliability for parameter research. To solve the above problems, based on a large number of construction parameters of a composite section subway in Shenzhen, this paper combined dimensionality reduction data with service analysis to optimize the selection process of shield tunneling parameters, and determined the total propulsion force, cutter head torque, cutter head speed, and advance rate as key tunneling parameters. Based on an LGBM algorithm and Bayesian optimization, the prediction model of key tunneling parameters of an earth pressure balance shield was established. The results showed that the average error of the LGBM model on the test set was 8.18%, the average error of the cutter head torque was 13.93%, the average error of the cutter head speed was 3.16%, and the average error of advance rate was 13.35%. Compared with the RF model, the prediction effect and the generalization on the test set were better. Therefore, an LGBM algorithm could be used as an effective prediction method for tunneling parameters in tunnel construction and provide guidance for the setting of tunneling parameters.

Dieses Werk wurde unter der Creative-Commons-Lizenz Namensnennung 4.0 International (CC-BY 4.0) veröffentlicht und darf unter den Lizenzbedinungen vervielfältigt, verbreitet, öffentlich zugänglich gemacht, sowie abgewandelt und bearbeitet werden. Dabei muss der Urheber bzw. Rechteinhaber genannt und die Lizenzbedingungen eingehalten werden.