Research Project GNN4GC

Transmission grids play a crucial role in the distribution of electricity between states, countries and continents and are the backbone of the electricity supply. They are becoming increasingly important due to increasing electrification which addresses the goal of an emission-free society. However, transmission system operators face challenges from decentrally generated renewable energy, electromobility and limited investment in new infrastructure projects. These factors lead to bottlenecks in the grid. To solve these, grid operators rely on expensive redispatching, where they have cut energy from other plants and still pay for it. A cost-effective alternative is to change the topology of the transmission grid. So far, this has hardly been used since the high number of switching points and switching options is difficult for human operators to keep track of. Here, digitisation and AI methods, especially Graph Neural Networks (GNNs), offer a solution. The GNN4GC project (Graph Neural Networks for Grid Control) aims to support grid operation by accelerating load flow calculations and evaluating and selecting suitable switching actions for the grid topology. GNNs are adapted to grid structures and thus represent a suitable method for managing the energy transition reliably and cost-effectively. By working with industry partners, the project opens up the possibility of bringing this method into application and thus overcoming challenges together.