An Innovative Holistic GIS-BIM and Artificial Intelligence Based Approach to Manage Mechanized Tunnelling: The Back-Analysis of the Budapest Metro Line4

Abstract

The current and continuously increasing demand for urban mobility implies introducing new sustainable and alternative systems to road transport. Where economic viability is established, metro lines are one of the most effective and least impactful solutions if the characteristics of the subsoil are appropriately considered and the construction phases are planned in such a way as to limit the induced ground deformations and not compromise the existing building stock. The excavation of tunnels in loose soils inevitably causes movements in the topsoil resulting in a combination of sagging and hogging, which in an urban environment must be controlled and minimized to avoid damage to the existing structures and infrastructure. Through the back-analysis of the Budapest (Hungary) Metro Line4, in this work, we propose an innovative tool where the design process is based on a GIS-BIM interaction, and the executive phase takes advantage of artificial neural networks capable of adjusting the design choices to the monitoring evidence. The environmental and geotechnical aspects are managed through the GIS Platform; then, 3D subsoil and structural models are developed following the BIM approach. After, the artificial neural network’s architecture is first constructed via a trial-and-error process which leads to selecting the best combination of input variables that better correlate to the measured volume loss. Then, real-time analysis is performed, and the transient effect is considered to simulate the excavation advance. The obtained results denote significant effectiveness in predicting the ground deformation and, thus, damage induced at the surface by mechanized excavation.