Digital twins can contribute to energy savings and sustainability goals by enabling detailed analysis and optimization of energy flows and operating strategies. However, the dynamic nature of digital twins, especially in the design phase and the transition to real-time operating conditions, requires an efficient and adaptable approach to creating and updating simulation models. This study presents a method for the automated creation of simulation models for digital twins that significantly reduces the model development effort while ensuring the consistency and currency of the digital twin. The method seamlessly integrates various aspects of model creation, from initial geometry generation to the final post-processing phase. The systematic organization and storage of data and metadata within the SIMULTAN meta-datamodel ensure consistency and currency across different process stages. This is demonstrated by a detailed simulation of Thermally Activated Building Systems (TABS) using Computational Fluid Dynamics (CFD).