Special Session

Machine Learning & Data-driven based TBM Tunnelling

 

Session Chair

Prof. Zixin Zhang, Tongji University

​

Co-chairs

Assoc. Prof. Xin Huang, Tongji University

Dr. Shuaifeng Wang, Tongji University

 

Though the TBM tunnelling technologies have been greatly improved and our experience of TBM tunnelling have been profoundly accumulated in the past decades, ensuring a fast and safe TBM tunnelling is still challenging. The enhancement of excavation efficiency and safety can be hampered by a number of extrinsic and intrinsic factors, including the complex operation parameters, great uncertainties in the ground formations and complicated TBM-ground interactions, etc.. A promising solution for this issue is to apply machine learning & data-driven based intelligent models, which abstract the complex physical or mechanical issues into a set of optimization problems. A number of such practices have been conducted on many aspects of TBM tunneling by resorting to a variety of intelligent algorithms, such as the conventional ML algorithm, random forest (RF), and deep learning algorithms, long short-term memory (LSTM) algorithms. Nevertheless, machine learning & data-driven based tunnelling is still an extensive area that has hardly been presented in a comprehensive and at the same time relatively simple way. This session provides a platform for scientific researchers and practical engineers to showcase their most recent advancements in the development as well application of intelligent models for TBM tunnelling. We also expect free and open discussions on the current difficulties/limitations of intelligent models in solving TBM tunneling issues, whereby future directions can be prospected.

​

This session plans to collect at least 7 high quality works in forms of either oral or poster presentations. Topics related to the scope of this SS include, but are not limited to, the application of machine learning & data-driven based models for:

• Predicting TBM operation parameters

• Predicting disc cutter wear/damage

• Predicting tunneling-induced ground responses

• Estimating strata composition ahead of excavation face

• Detecting unfavorable geological objects

​