Enhanced differencing and merging of IFC data by processing spatial, semantic and relational model aspects

Building Information Modeling promotes collaborative work in which models are edited and enhanced in parallel by domain experts. In this environment, revised and extended datasets have to be combined to an overall database. To perform this merging, duplicates have to be identified. This paper describes a novel, comprehensive approach for detecting equivalences in datasets of the Industry Foundation Classes (IFC), an open and full-fledged schema for building models. In contrast to available methods, the presented approach is independent of object identifiers. Instead, the detection of duplicates is performed on entity level considering spatial, semantic and relational aspects of the IFC data.