From BIM databases to Modelica - Automated simulations of heating systems

Abstract

Detailed simulations of HVAC systems play a crucial role in creating 1:1 digital twins of buildings and their systems. In particular, detailed models of hydronic systems are essential for fault detection of building services and control optimization. However, modeling HVAC systems is labour intensive due to the components and connections that one must create based on drawings or models. Creating the HVAC simulation models from BIM data eases the modelling burden, simplifying the creation of digital twins. Straight-forward HVAC simulations can aid the design process. Instead of prescriptive design based on the worst-case conditions, simulations enable performance-based design with partial-loads and dynamic behaviour. This paper presents a preliminary tool using BIM data to create and simulate models of heating systems. The tool uses a central BIM data platform with a dedicated data format – defining components and their relations in a database. Python scripts apply model templates to create heating system models in the Modelica language. The tool simulates the models in Dymola, while Python scripts read and parse the results to the database for visualization and analysis. The tool efficiently simulated a small heating system and obtained results for the return temperatures of several loops.