Fraunhofer-Gesellschaft

Publica

Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Artificial Neural Networks (ANN) for the Prediction of Local Outside Temperatures and Solar Yields

 
: Kramer, W.; Bitterling, M.

:
Fulltext ()

International Solar Energy Society -ISES-; International Energy Agency -IEA-, Paris:
SWC 2017/SHC 2017, ISES Solar World Congress, IEA SHC Solar Heating and Cooling Conference 2017. Online resource : 29 October - 02 November 2017, Abu Dhabi
Abu Dhabi, 2017
http://proceedings.ises.org/?mode=list&conference=swc2017
6 pp.
International Conference on Solar Heating and Cooling for Buildings and Industry (SHC) <2017, Abu Dhabi>
Solar World Congress (SWC) <2017, Abu Dhabi>
English
Conference Paper, Electronic Publication
Fraunhofer ISE ()
Thermisches System; Gebäudetechnik; Thermischer Speicher; Gebrauchsdaueranalyse; Predictive Control; Solarthermie; Gebäudeenergietechnik; Thermische Anlagentechnik; Wärme- und Kälteversorgung

Abstract
Artificial Neural Networks (ANN) are the basis of a new intelligent control concept for residential heating systems developed at Fraunhofer ISE. This artificial intelligence based concept is able to predict certain input data like local weather, solar thermal yield, thermal behavior of a building and capacity level of a thermal storage based on measured data without using any simulation tool. This allows to improve energy efficiency and simplicity of control devices at the same time by considering the above mentioned predicted data. The current paper shows results of prediction by ANN for local outside temperature and solar yields of a solar thermal installation. The focus of the prediction method is not highest accuracy but simplicity of application, enabling low cost model predictive control. Accuracy of the local outside temperature prediction is doubled compared to a pure weather forecast data for the tested location. Solar yield prediction is also quite in line with real measurement data at the location investigated. The achieved prediction quality is reasonable and promising for improved heating system control.

: http://publica.fraunhofer.de/documents/N-525492.html