The energy sector is facing rapid changes in lieu of transitioning toward clean, renewable sources. The growing share of volatile, fluctuating renewable energies and the increasing share of cross-regional energy exchange, have recently resulted in increasing grid congestion, especially when it comes to ensuring a secure supply of energy.

Last week during the State Innovation Award Gala, the JKU and a start-up company, enliteAI, were presented with a special award, the VERENA - powered by Verbund.

The award-winning project is a globally unique method featuring reinforcement learning agents that can determine the ideal switching states for an entire power grid (thereby avoiding blackout situations), and apply ‘intelligent flow control’ to reduce costs and CO² emissions during short-term capacity measures.

Stromnetz 4.0 was developed in partnership with the Institute for Machine Learning and the Ellis Unit Linz and in 2022, the start-up company won first place at the global "I2rpn Challenge", a call published at international conferences by the French network operator RTE. The underlying reinforcement learning methodology was accepted as a workshop paper during the Neurrips'22 Conference and presented at a workshop.

See: https://www.enlite.ai/solutions/energy, opens an external URL in a new window to learn more, as well as read the accompanying paper on Arxiv: https://arxiv.org/abs/2211.05612, opens an external URL in a new window