BIM-based object mapping using invariant signatures of AEC objects

Building Information Modeling (BIM) can facilitate building engineering and performance analysis from the early design phase. Although BIM provides engineers with building objects and their information that can be used in a variety of platforms for creating analytical models, the shared objects have often lost track of their identification in different models during model creation, design iterations, and model transformation processes. To address this, mapping algorithms that can interface BIM and analytical models (e.g., building energy models (BEM), structural analysis models) were developed in this paper by automatically mapping objects based on their invariant signatures. An experiment was conducted on mapping space objects between BIM and BEM following the proposed framework. The developed algorithm was tested on a 4-story office building model with 82 spaces. The results showed that the algorithm achieved 90% precision and 90% recall for mapping space objects. A 4.88% improvement in the accuracy of energy simulation results was achieved through the algorithmic space mapping compared to that without space mapping. This research will contribute to filling informational gaps between BIM and analytical models. The proposed framework can also be used to further expand BIM inter -operability support by providing a new data-driven approach for developing transformation algorithms.