Automated classification and coding for BIM components based on applicable big data

Abstract

For the application of BIM technology, BIM models play a vital role in transferring and sharing building information. BIM models represent the building information through the attributes of components, such as walls and slabs, and the relationship between these components. Among all the attributes of components in BIM models, the classification and coding attribute is essential to retrieve the building information in a quick way. However, in practice, the BIM models that are prepared by using a BIM authoring tool cannot ensure complete and correct classification and coding attribute when they are transferred into the format of another tool. Besides, BIM models prepared by using 3D reconstruction technology also lack the classification and coding attribute. Missing or incorrect classification and coding attribute of BIM components impedes the fully exploitation of BIM model greatly. To solve the problem, this paper proposes an automated classification and coding method for BIM components based on a batch of BIM models with components labeled with key features and type, which can be obtained from the big data of the BIM models that have been correctly used. In the method, the association rule mining algorithm is used to establish the classification and coding rules for BIM components based on the labeled component data set. Then, for any BIM component, an algorithm based on the credibility reasoning approach is used to execute the rules and obtain its classification and coding attribute. In this way, the classification and coding attribute can be determined for any BIM component according to any given standard. The method is validated by developing a prototype based on Autodesk Revit and by using a batch of BIM models from structural design, and 92.7% precision is achieved in the test case. This method contributes to the quick classification of BIM components.