AI-enhanced data workflows for road infrastructure BIM: a Python–VPL framework enabling 4D–7D integration

Abstract

The transition of road projects from a purely geometric, two-dimensional or three-dimensional configuration to one rich in information for sustainability-oriented management requires a multi-dimensional approach, i.e., exploiting the full potential of Building Information Modelling methodology. This work proposes an AI-backed BIM workflow which, by adopting the solutions offered by an AI agent, integrates Python scripts to automate analyses and calculations related to time-, cost-, and sustainability-related assessments. The methodology was validated on a case study of a suburban infrastructure, involving a road intersection and a hydraulic underpass. The IT tools used are those specified in the tender documentation, in particular Autodesk Civil 3D, Revit, and Navisworks software. In this environment, Dynamo was used to insert Python scripts into an environment that supports visual programming language. The result is models exported in open IFC 4.3 format. The approach adopted suggests that AI can be successfully exploited to increase automation levels in BIM processes. This leads to a reduction in human error, shorter timescales, and therefore lower costs, thanks to the acceleration of iterative processes and the reduction of repetitive tasks. Its application to computations and analyses supporting cost calculation, scheduling construction and work phases, and calculation of carbon footprint as an index of sustainability opens up to increasingly relevant uses in terms of environmental sustainability and project quality. The results suggest that the integration of AI, visual and textual programming methods, and established BIM-Authoring software can lead to innovative sustainability techniques and digital twin approaches in infrastructure management.