Image Copyright: fotolia.com/thomaslerchphoto

The Fraunhofer Institute (IEE) in Kassel has developed a new forecasting system which is based on artificial intelligence.

© fotolia.com/thomaslerchphoto

Germany’s power grids were set up to run on nuclear, coal and oil, but in the process of Energy Transition they will be converted to renewable energy sources such as wind and solar. These new smart grids must be better automated and digitally controlled. A new project called “iAutomate” led by the Technical University of Dortmund’s Institute for Energy Systems, Energy Efficiency and Energy Economics (ie³) has set out to thoroughly investigate and test how in practice these grids can be made “smart”.

Starting in the town of Korbach, Hesse, in collaboration with the local energy supplier Energie Waldeck-Frankenberg, the team installed special measuring devices called Energy & Power Protocolling Equipment in the local network transformers, which were equipped with specially-developed software for the purpose.

Firstly, they simulated grid operations in the laboratory before conducting practical tests on site. The algorithms observed network conditions – looking for bottlenecks and shortages – and sent the data to mission control where it was manually checked over. Automated controls will be vital in the future to avoid scenarios like the grid becoming maxed-out when everyone charges their e-vehicles at the end of the day.

The field test showed for the first time that it is possible to measure energy flows and voltage throughout an entire network (previously, visibility was limited to individual network transformers). “With our work, we have managed to ensure that the grid can now be better controlled”, says Dominik Hilbrich, head of the research group at the TU in an article on Energiesystem-Forschung, “This is of great importance because the modern energy system is becoming more and more complex due to the increasing demands”.

The software uses flexible system architecture which means it can be universally adapted for different networks and tasks.