The refurbishment of existing buildings is one important aspect of current economic and environmental policy at national and international level. The automated simulation of different renovation scenarios offers extensive potential for the identification and planning of optimal renovation measures. However, simulations of existing buildings in particular require specialized building information modeling (BIM) with a very high level of detail. At present, this has to be created and calibrated manually, as the data is usually incomplete.

The goal of the ZIM-funded project KI-Sanierung is to develop an AI-based system that automates the digitization and calibration of BIM models for existing buildings with limited data availability. For this purpose, different AI methods (machine learning, planning, evolutionary algorithms) are to be used and integrated into a suitable workflow. This project is implemented within the framework of the ZIM funding of the BMWi together with the project partners Metabuild and Deimel Oelschläger Architekten Partnerschaft.