Towards a data-driven assistance system for operating segment erectors in tunnel boring machines

In tunnel construction, heavy machinery like tunnel boring machines are used to increase the safety and efficiency of tunneling projects. Despite the fact that these machines are equipped with a variety of different sensors, human workers still operate both the advance and ring building phases of the machine. Due to the high complexity of handling segment erectors, the ring building phase in particular has a lot of room for improvement. In this work, a concept of a data-driven assistance system to support operators in the ring building process is proposed. Three approaches to extract human expertise from operational data of tunnel boring machines using machine learning and explainable artificial intelligence are presented and discussed. Furthermore, preliminary data preprocessing results verify that our approach of determining ring building times is more accurate compared to on-site measurements. This contribution highlights the untouched potential of ring building optimization using machine learning techniques based on real-world data.