NEBT Project: The Application of Artificial Intelligence (AI) to Improve TBM Operations - RETC2021

The application of artificial intelligence (AI) to improve tunnel boring machine (TBM) operations requires a significant amount of information to train robust performance prediction models, e.g., for tunneling-induced settlement, productivity, tool wear, etc. This paper presents the AI framework applied to the North-East Boundary Tunnel (NEBT), the largest component of the Clean Rivers Project in Washington, DC. The paper focuses on required elements for real-time AI use to estimate tunneling-induced ground and building deformation and cutter tool wear. Keys to the framework include defining the necessary quantities of data required as well as the process of windowed training, validation and model testing. To date, the models developed for ground deformation and tool wear show satisfactory performance when compared to observed field observations during tunneling.