Data4Grid Challenge: Fraunhofer IEE Consumption Forecasts receive an Award
Dominik Jost, Jonas Koch and Dominik Beinert, scientists at the Fraunhofer Institute for Energy Economics and Energy System Technology IEE, have won the Data4Grid Challenge "AI-supported consumption forecasts based on smart meter data". The competition was organized by the German Energy Agency (dena) and umlaut. The results and award ceremony presentation took place at the Future Energy Lab in Berlin on May 4, 2022.
Transparent information has a key role in operating distribution grids and thus enabling the energy transition towards high penetration of renewables. In the power grid of the future, it will become increasingly important to reliably forecast the power consumption of all players and guarantee a stable grid status even during fluctuating power generation caused, for example, by PV Systems. Growing electro-mobility and the increasing use of heat pumps, for instance, are reflected in a significant change in consumption behavior on the low-voltage level.
The digitization of distribution networks - for example in the form of smart meters - offers great opportunities to create transparency on the medium and low voltage levels and to enable distribution network operators to operate their networks efficiently and securely. In addition to generation forecasts for distributed generators, this also requires accurate consumption forecasts.
Reliable forecasts of generation and consumption are the scientific focus at Fraunhofer IEE in the business unit of Energy Meteorological Information Systems. Now, Fraunhofer's expertise in consumption forecasts has been awarded first prize in the Data4Grid Challenge "AI-supported consumption forecasts based on smart meter data". The Fraunhofer IEE team, Dominik Jost, Jonas Koch and Dominik Beinert, prevailed against five competitors and received prize money of 4,500 EUR. The results of the three challenges were presented as a pitch at dena's Future Energy Lab in Berlin.
As part of the challenge, Fraunhofer scientists analyzed smart meter measurements and used them to generate AI-based forecasts for individual consumers and clusters for the next 24 hours. Smart meters are continuously increasing the amount of data available and therefore provide a reliable basis for consumption forecasts. In the future, they will provide the foundation for automated grid control in combination with generation forecasts from decentralized power suppliers.
In addition to its extensive expertise in creating precise power forecasts, Fraunhofer IEE's many years of experience in working in grid operation also contributed to its success in the challenge. The Kompetenzzentrum Kognitive Energiesysteme (K-ES), which aims to link artificial intelligence and the energy industry, supported the participation.
Furthermore, the Fraunhofer IEE spin-off retoflow and the team Dr. Leon Thurner, Simon Ruben Drauz and Marcel Dipp won the first prize in the challenge "Evaluation of relevant measuring points to increase network transparency in low-voltage networks".
As part of the challenge, an expert report on critical processes in the power grid was drafted. The report was conducted by Fraunhofer IEE under the leadership of Kristina Jurczyk and will be published by dena in the course of the summer.