The quality of a tunnel depends on the quality of its segment installation. And the quality of the segment installation depends on how precise the tunnel boring machine (TBM) trajectory moves along the tunnel design alignment (TDA). Therefore, the quality of a tunnel strongly depends on the skill of operators who steer the TBM.
However, steering a TBM is a difficult task. It is strongly influenced by the uncertain ground conditions. And the TBM movements are the product of interactions among the geology, the operators control decision, and TBM shield behaviors. To steer a TBM, the operators have to continuously make real-time interpretations, judgments, and decisions based on signals produced by TBM sensors and the guidance systems. It also consists of trial and error processes. This study investigates the applicability of machine learning algorithms to be employed as guidance and control systems to better navigate the TBM.
Publications
- Apoji, D., Fujita, Y., Soga, K., 2020, “Toward Autonomous TBM: Preliminary Study of Dynamic Prediction of Shield Position for TBM Trajectory Control”, UC Berkeley Geosystems Engineering Annual Research Symposium (poster).