ECPPM 2022 - eWork and eBusiness in Architecture, Engineering and Construction 2022
Raimar J. Scherer, Sujesh F. Sujan, Eilif Hjelseth
Abstract
A method and structure for architectural datasets specifically designed for the analysis, sorting, and ultimately reusing of building elements is proposed. Four different methods of parsing data from real-life projects using their building information models (BIM) for integration into a machine learning (ML) model were evaluated. As ML integration is becoming more important in the Architectural Engineering and Construction (AEC) industry, we see an increasing demand for high quality datasets. Four different methods and file formats were benchmarked, focusing on read and write-speeds for converting architectural BIM into datasets to be used in ML. Our results show that the current way of storing our projects in Industry Foundation Classes (IFC) is not optimal for the development and integration of new Artificial Intelligence (AI) assisted tools. This paper provides alternative methods and storage solutions for both developing new datasets internally and also for future work in creating a common federated learning setting for the AEC industry.