Automating HVAC enrichment in building energy models from imperfect BIM data through topology completeness validation

Knowledge of a building’s HVAC topological information can be especially valuable in developing a first-principles energy model, which is particularly beneficial for building design, operational planning, and management. It is critical in facilitating the Building Information Modelling to Building Energy Models (BIM2BEM) process, where the BIM model serves as the primary data source. The imperfections in BIM data, such as missing connections or elements, pose challenges in automatically and effectively extracting HVAC-related topological information. To facilitate the BIM2BEM process when dealing with imperfect BIM data, this work presents an HVAC-enriched BEM generation framework that addresses varying levels of HVAC topology completeness. It begins with establishing an initial HVAC topology by utilising ontologies to represent essential components and abstract their connections. A workflow to ensure the completeness of this topology is employed, utilising a hierarchical ruleset that includes Terminal-Zone, Air Loops, and Water Loops. Subsequently, a series of queries is constructed to extract relevant information from the HVAC topology, enabling the automatic population of BEM models. This framework yields three distinct BEM models, i.e., Ideal Load, Partial Match, and Perfect Match, each representing different levels of HVAC topology completeness. A case study involving a newly constructed building with a complex HVAC system demonstrates the effectiveness of the proposed framework by verifying the simulated energy indicators against the building energy bills. Finally, a comparative analysis highlights the impact of varying HVAC topology completeness on building performance simulations and configuration design. The proposed framework offers a scalable solution to bridge the gap in HVAC information extraction from BIM to BEM, with generalisation capabilities across varying levels of data quality.