AI-powered weather forecasts are improving predictions for smart grids' energy outputs
National Grid is using artificial intelligence (AI) and machine learning to help predict how much energy the United Kingdom will reap from turbines and solar panels when the wind is blowing or the sun is shining.
Thanks to a new partnership with the Alan Turing Institute, National Grid Electricity System Operator (ESO) announced it has developed new AI prediction models that have improved solar forecasting by one-third.
Knowing how much power will be flowing into the grid on any given day is becoming increasingly crucial as the proportion of intermittent renewable power serving the grid goes up.
Rob Rome, commercial operations manager at the ESO, said the new forecast models means the power system can become much more efficient at managing supply and demand.
"Improved solar forecasts will help us run the system more efficiently, ultimately meaning lower bills for consumers," he said. "It will also enable more solar capacity to be connected and utilized, helping us to achieve our 2025 ambition to be able to operate a zero-carbon electricity system."
National Grid worked with researchers and doctoral students at the Institute to develop the improved forecasting models. The new system combines information including temperature data, solar irradiation data and historic weather data to reach an output generation figure, which is then tested against 80 weather forecasts to give an energy generation forecast.
The project is similar to a collaboration between Google and AI firm Deep Mind, which is also using AI technologies to better predict energy output from wind farms.
Google revealed in February it can predict output from its wind farms 36 hours in advance, allowing the tech giant to bid to serve power to the grid ahead of time, a service which commands a higher power price.