The significant increase in the capacity of renewable energy in the world has led to numerous challenges when it comes to the management of the electric grid. Since the production of electricity from renewable sources is highly uncertain, while the increase in the number of electric vehicles and distributed energy production systems introduce changes in demand, traditional methods of grid management struggle to cope with these challenges in real time.
Inspired by the emerging problem, researchers from the University of Virginia have developed an artificial intelligence model that increases the reliability and efficiency of networks under conditions of increasing uncertainty, especially in the context of increased use of green energy and electric vehicles.
The new model is based on multi-fidelity graphical neural networks (GNNs) – a type of artificial intelligence designed to improve energy flow analysis. A “multi-fidelity” approach allows an AI model to use large amounts of lower-quality (low-fidelity) data, while simultaneously using smaller amounts of highly accurate (high-fidelity) data.
Low-fidelity data is often used in larger quantities because it is more readily available and cheaper to collect, although it is not as accurate or complete as high-fidelity data, which is more accurate but usually more expensive and complex to collect. Low-fidelity data allows the model to quickly learn general patterns, while a smaller amount of high-fidelity data helps fine-tune the model to achieve greater accuracy.
By applying this model, the system can make real-time decisions and adapt to changes, such as power line failures, optimizing the production and distribution of energy from different sources in accordance with changing network conditions. The new AI model integrates both detailed and simplified simulations to optimize solutions within seconds, improving network performance even under unpredictable conditions.
“Renewables and electric vehicles are changing the landscape, and we need smarter solutions to manage the grid,” said Negin Alemazkur, assistant professor and lead researcher on the project. “Our model helps make quick and reliable decisions, even under unpredictable conditions.”
This model brings key advantages: it requires less computing power, it is more accurate and it is resistant to changes in the network, such as line failures. It represents a step towards a more stable energy future, facilitating the integration of renewable sources and providing the foundation for more reliable electricity grids.
Energy portal
Source: energetskiportal.rs
