AI-based methods for optimising the operation and expansion of distribution grids

"The way to the energy transition is through the power grid". Increased decentralized, renewable power generation is already posing major challenges for distribution grids. There is a threat of rejections of new distributed generation or curtailment of generation by grid operators.

How will this situation look with regard to the goals of expanding renewable energies in Austria?

Until now, there was hardly any need for grid operators to know the current operating status of the low-voltage networks. Extensive knowledge was available for the operation of the grids, and supply interruptions were rare in Austria. The expansion of distributed, renewable generation has changed this situation.

Advancing digitalization, or artificial intelligence, offers a great opportunity to provide knowledge about the current state of the distribution grid, to better predict energy generation, and to use new methods for grid monitoring and grid planning.

The AI4GriDs project focuses on small grid operators (in Austria, only 11 of the 122 DSOs have more than 100,000 customers), as they in particular are currently unable to provide the data required for AI training and operation. This is mainly due to a lack of sensors, communication infrastructure and data management.

The aim is to provide grid operators with the necessary information to prepare for data collection so that AI methods can be used for RES forecasting as well as for network planning and network operation. The AI4GriDs project researches the methods, provides experimental evidence of their applicability and forms the basis for a demonstration as part of a follow-up project.

• GeoSphere Austria (vormals Zentralanstalt für Meteorologie und Geodynamik – ZAMG)
• EnliteAI GmbH
• Stadtwerke Kapfenberg GmbH
• Stadtwerke Mürzzuschlag Gesellschaft m.b.H.