In China, the construction of the Yangcun Tunnel on a 350 km/h high-speed rail (HSR) line in the Hubei Province was the first time that an AI system was used to select the tunnelling method. A deep learning model, trained with data from 251 HSR tunnels and 1,700 construction segments, analysed the geological parameters and divided the tunnel into micro-sectors to determine the best construction method for each segment, including the CD method, drill and blast or excavator. The AI achieved a prediction accuracy of 89.41% and was checked by experienced engineers and integrated into the BIM. This was the first time in the world that artificial intelligence (AI) was used to select and approve the main construction of an HSR tunnel. The procedure is considered a milestone for the use of AI decision-making processes in large-scale infrastructure construction and a reference model for future railway projects.
