AI-Based Architectural, Mechanical, Electrical and Plumbing BIM Object Classification Model and Semantic Enrichment Framework
Enobong Archibong
Department of Computer Engineering, University of Uyo, Akwa Ibom State, Nigeria.
Bliss Stephen
Department of Computer Engineering, University of Uyo, Akwa Ibom State, Nigeria.
Michael Esu
Department of Computer Engineering, University of Uyo, Akwa Ibom State, Nigeria.
Philip Asuquo *
Department of Computer Engineering, University of Uyo, Akwa Ibom State, Nigeria.
*Author to whom correspondence should be addressed.
Abstract
Building Information Modelling (BIM) is increasingly being adopted across the architecture, engineering, and construction industries for digitally simulating and managing infrastructure projects. Despite the growth in BIM utilisation, challenges persist in the classification and semantic enrichment of architectural, mechanical, electrical, and plumbing (MEP) objects, critical components in infrastructure modelling. While some studies have addressed object classification or semantic enrichment independently, there is limited research on integrating both, particularly for MEP components where semantic clarity and interoperability are essential for effective cross-disciplinary stakeholder collaboration. This paper introduces a novel AI-based framework for architectural MEP BIM object classification and semantic enrichment, incorporating multiple deep learning components. The proposed system leverages 3D Convolutional Neural Networks (CNN) for spatial feature extraction, Graph Neural Network Transformers for capturing relational features, and a CNN-based feature fusion model
Keywords: Artificial Intelligence (AI), Building Information Modeling (BIM), AEC, MEP, deep learning, semantic enrichment