An Ontology-Based Data Model to Create Virtual Training Environments for Construction Safety Using BIM and Digital Twins

Accident rates in the construction industry remain high. Recent research introduced serious games based on extended reality and proved increased safety awareness among the trainees. However, creating such training scenarios requires resources across many domains. This work aims to validate that a Digital Twin for Construction (DTC) provides the necessary data to generate a virtual training environment automatically. We propose a novel ontology-based data model integrating a serious game into a DTC. The data model includes the project intent information, safety regulations, and status knowledge to update the learning environment dynamically. A prototype is developed in Unity based on two test projects (building and infrastructure) that contain several typical construction hazards. The case study indicates the data model's viability for generating a virtual environment with little means based on the DTC. This work further suggests value once the underlying data sources for the model are connected and of high quality.