To support local grid stability, the complex system requires the use of AI to estimate the current grid state, to predict electricity generation from renewable energy sources and to support grid system services that use battery charging and discharging. A complex system in which the components of a future power grid are interconnected is susceptible to faults, especially if the components are controlled in a central location. In order for AI to work safely, a distributed AI approach is being investigated. This distributed AI approach uses cognitive edge computing for efficient control, to reduce resources and to increase data security. The principle is to run applications as close as possible to the data sources. AI4DG’s goal is to research and develop a decentralized AI platform for a safe and autonomous control of the distribution grid with a high proportion of renewable energies.
Project partners:
Germany: Bielefeld University, Bielefeld University of Applied Sciences, Westfalen Weser Netz GmbH
France: Université Grenoble Alpes, Atos Worldgrid
Project funding:
German-French cooperation, Federal Ministry of Education and Research (BMBF)