Data mining, machine learning, feedback and feedforward – Energy efficiency through usage-centred building systems

Runtime: 3.5 Years

Start: 7/2019

End: 12/2021

Project partners:

  • ABB AG, Building Automation
  • Bayern Facility Management GmbH
  • Fraunhofer IBP, Department Indoor Climate
  • Karlsruher Institute of Technology, Building Science Group
  • RWTH Aachen, Institute of Energy Efficiency and Sustainable Building

Funding body: BMWi - Bundesministerium für Wirtschaft und Technologie, funding reference: 03EN1002B

  BMWi Copyright: BMWi

Numerous studies on the energetic performance of buildings show that users have a significant influence on energy consumption. At the same time, inadequate knowledge of user behaviour in practice often leads to a large difference between the energy indicators predicted during the design phase and those actually measured during operation. The aim of this project is therefore to reduce this performance gap by systematically opening up and optimizing the data usage process chain in order to enable more reliable forecasts for building operation and ensure high energy efficiency.

For this purpose, the use of equipment and technical facilities as well as comfort related interactions of the users are recorded more precisely, processed and made available in the form of models for planning tools and systems for operational management. Intelligent sensor technology, data mining, machine learning or predictive analytics are used to develop more efficient analysis and modelling methods in order to improve building performance and obtain better assessments of user behaviour. The digital twin will also play a central role here as a cyber-physical image of the real devices and buildings, similar to what is clearly foreseeable in Industry 4.0. Building on this, services for building users and operators will be developed, for which questions of information processing and communication will be in the foreground. At the user level, this is manifested by the presentation of essential indoor climate and other environmental parameters together with decision aids for the optimization of user group-specific or individual comfort. On the operator level, on the other hand, services are provided through reports on the development of energy demand in various sectors and recommendations for action derived from these reports. With regard to a more comprehensive understanding of user behaviour and its interaction with building equipment, the perception of and reaction to multiple and partly interacting environmental influences will also be investigated experimentally.

The developed solutions will be tested and evaluated in demonstration projects. In addition, a substantial contribution will be made to the IEA EBC Annex 79 "Occupant behaviour-centric building design and operation", which was launched in autumn 2018 and in which the project participants are involved in a leading role.