To control a tunnel boring machine (TBM), operators have to continuously make real-time interpretations, judgments, and decisions based on signals produced by TBM sensors. For example, in soft ground tunneling using an earth pressure balance shield machine (EPBM), the operators have to interpret the encountered geologic conditions, decide the excavation control parameters, inject foam to condition the encountered ground, balance the chamber soil pressure, and navigate the shield to achieve correct tunnel position.
For humans, continuously making real-time interpretations, judgments, and decisions in various control tasks based on numerous streaming sensor data is not straightforward and may lead to inconsistent results. A more systematic approach is required to utilize TBM operation data. This study implements Bayesian networks (BN) and structure learning algorithms to model the interactions among excavation features of an earth pressure balance shield machine (EPBM).
Publications
- Apoji, D., Fujita, Y., Soga, K., “Exploring Interactions among EPBM Features using Bayesian Networks”, submitted to Word Tunneling Congress 2022.