Semantic Enrichment of As-is BIMs for Building Energy Simulation

Abstract

Recent efforts on automatic 3D modeling of existing buildings often result in semantically poor as-is building information models (BIMs). Such a BIM usually consists of a set of objectified surfaces characterized by the building element types they represent and 3D surface geometries. It cannot be directly used as the geometry input of building energy modeling (BEM) as the key concepts of second-level space boundaries (SBs) are missing. This paper proposes a semantic enrichment approach for automatically adding such semantic concepts inferred from the semantically poor as-is BIM. The output of this approach is a corresponding IFC BIM with second-level SBs, which can be further used by various energy simulation tools. Preliminary experimental results using a building surface model demonstrate the performance of the proposed approach.