An Ontology-Driven Framework for Intelligent Embodied Carbon Decision-Support in Metro Station Design

Abstract: Conventional element-centric BIM-LCA integration omits temporary works and faces semantic barriers, causing significant carbon underreporting. We propose a dynamic ontology-driven framework to overcome this. Its semantic core is a novel System-Element-Process-Resource-Carbon ontology, shifting assessment from static elements to context-aware construction processes. An automated pipeline transforms OpenBIM data into a carbon knowledge graph, capturing permanent and temporary emissions. A neuro-symbolic multi-agent system enables natural language querying and scenario simulation for interactive decision support. The framework ensures more complete and automated carbon accounting and establishes a semantic foundation for data continuity across the project lifecycle.