Generating second-level space boundaries from large-scale IFC-compliant building information models using multiple geometry representations

Automatically extracting information needed for building energy modeling (BEM) from an IFC BIM can significantly improve efficiency and accuracy in preparing BEM inputs. Second-level space boundaries (SBs) define geometric data required for BEM, but they are often missing or incorrectly defined in IFC models. To address limitations in existing algorithms when generating second-level SBs of complex buildings, this study presents a novel approach that computes geometric and semantic information of second-level SBs using multiple geometry representations of building elements and spaces in an IFC model. Advantages of the approach include: it has relatively loose requirements on the quality of input IFC models; it can process building objects with common curved shapes; it improves computing efficiency when processing large-scale building models by employing three types of geometric representations. The approach was evaluated with two large-scale real-world building models, and the results show that second-level SBs are generated accurately and highly efficiently.