Self-learning, model-based method to identify process models of the system components.
Supply systems in buildings compose of different, individually controlled system components. Despite interactions between those components, no holistic adjustment of control parameters takes place during the commissioning phase.
In this project a self-learning, model-based method is developed to identify process models of the system components using measurement data recorded during operation. The identified process models are used to continuously derive optimal control parameters for the current operating point, taking interactions between system componentes into account. Applying this method periodically allows the consideration of changing operating conditions and different user behaviour. The application is developed at our Hardware-in-the-loop test bench and demonstrated in field tests to show the feasibility under practical operating conditions.
Funding

Project term | November 2016 to November 2020 |
Funding code | 03ET1447A |
Project management | Yannick Fürst |