DataFEE – Data mining, machine learning, feedback and feedforward – Energy efficiency through needs-based building systems
The energy consumption of buildings is strongly influenced by occupant behavior. Oftentimes, a considerable performance gap between predicted and measured energy performance indicators can be observed – mainly due to a lack of understanding of occupants’ behaviors and the inability to predict it fairly well. By exploiting and optimizing the process chain of data usage, we aim to reduce this performance gap. In combination with a digital twin, a cyber-physical representation of equipment and buildings, tools such as Data Mining, Machine Learning, and Predictive Analytics will be employed. The description and prediction of occupant behavior will be improved by this approach, for example, by taking interacting environmental influences more strongly into account. Occupants as well as facility management will benefit from this type of data processing and feedback. This includes, for example, the display of relevant indoor climate indices or the provision of energy performance indicators accompanied by appropriate action recommendations. Ultimately, this will help to increase occupant comfort and make the operation of buildings more efficient. The developed models and algorithms will be tested and evaluated in demonstrator projects. Last but not least, we will contribute to IEA EBC Annex 79 Occupant behaviour-centric building design and operation.
Project term:
07/2019 - 06/2023
Project partner:
- Karlsruher Institut für Technologie, Fachgebiet Bauphysik und Technischer Ausbau
- RWTH Aachen, Lehrstuhl Gebäude- und Raumklimatechnik EBC
- Fraunhofer Institut für Bauphysik IBP, Abteilung Raumklima
- ABB AG, Corporate Research Center Germany
- Bayern Facility Management GmbH
Website:
Funding/Client:
BMWi, Förderkennzeichen 03EN1002B