National Grid Electricity System Operator (ESO) has launched a new project alongside the Smith Institute to develop an innovative approach to forecasting day-ahead reserve requirements.
The ESO said its Dynamic Reserve Setting project will help it to better understand uncertainties in its forecasting data and set reserve levels more accurately, limiting the need to keep fossil fuel plant running as backup.
Reserve levels are currently determined based on a number of factors including expected electricity demand at different times of the day and week, historical generation, forecast renewable generation and likely forecasting errors.
The ESO said these levels could potentially be optimised by linking forecasting errors to weather-driven effects on the system and buying reserve in day-ahead timescales. The Smith Institute has been enlisted to develop a proof-of-concept machine learning model that uses variables such as temperature and wind forecast data to create more accurate predictions of forecast errors to account for the variability of day-to-day weather.
Isabelle Haigh, head of national control for National Grid ESO, said: “As more clean energy connects to Britain’s electricity system, the network is becoming more challenging to operate. The more confidence and certainty we have in our forecasts, the more efficiently and securely we’ll be able to balance the country’s supply and demand day to day, minute by minute.
“Innovative developments like this are crucial if we’re to realise our zero-carbon ambition. Collaboration is key and we look forward to working with the Smith Institute to develop our processes.”
Rachael Warrington, executive mathematical consultant at the Smith Institute, said: “I’m really excited to be part of creating a new approach that could make a big difference in the energy industry. The way it could help ESO move towards carbon net-zero is also a real motivation.”
The project, which is being supported with funding from Ofgem’s Network Innovation Allowance, is scheduled to run for approximately 12 months, with the initial proof of concept expected in November 2021. If successful, the algorithm will be integrated into the ESO’s systems as part of a live trial.