Automated inference of context-specific hazards in construction using BIM and Ontology

To address the high rate of workplace accidents in the construction industry, this paper proposed an automated hazard identification process using building information modeling (BIM) and ontology. In South Korea, legislation mandates risk assessment and safety documentation to prevent construction accidents by identifying potential hazards. Current methods rely on the experience of personnel, which limits hazard recognition. The proposed approach leverages BIM to automatically infer construction methods, tasks, tools, and materials, identifying related hazards and mitigation measures through ontology. Validation experiments focused on waterproofing work revealed alignment between inferred risks and expected outcomes. By comparing the ontology-derived risk factors with those identified by safety managers, this study confirmed the consistency and adequacy of the ontology. The method improves accuracy, efficiency, and consistency in hazard identification in various construction projects. • Process for inferring construction methods and hazards using minimal BIM data. • The process links works, tools, and materials for context-specific hazard identification. • Validation with waterproofing work revealed improved consistency and reduced omissions. • The method supports safety planning from design to construction through automation. • This approach advances hazard inference beyond static mapping to dynamic automation .