Semi-automated code compliance checking based on semantic reasoning

Automatising the process of verifying whether the design of a building conforms to a certain regulation is an important task to save time and money. This paper presents SmartNorms4BIM , a building regulatory compliance-checking system that facilitates the automatic evaluation of Building Information Modelling (BIM) models. The system’s architecture follows Eastman’s four stages: translation of norms into a machine-readable format, BIM model preparation, execution of the verification process, and report generation. Our proposal uses Semantic Web technologies (OWL 2 ontologies and semantic reasoners) to represent the knowledge and detect violations of the norms automatically, Large Language Models (LLMs) to interpret and transform building norms into OWL 2 ontologies, fuzzy logic to manage vagueness in the norms, and specific algorithms to reason with geometric knowledge. A Java desktop prototype was developed to determine whether a BIM model complies with a given set of norms. To validate our approach, two real-world use cases were studied.