National Grid ESO - electricity margins

Dynamic Reserve Setting innovation project seeks to reduce balancing costs

We’re highlighting a project from our Innovation Annual Summary, which reduces consumer costs and utilises zero carbon electricity generation – Dynamic Reserve Setting.

What is Dynamic Reserve Setting?

To keep Great Britain’s electricity network stable, our control room must maintain a balance between supply and demand at all times. One of the ways we do this is by utilising reserves of power, which are procured in advance and remain on standby until needed, based on statistical analysis of historical generation, and forecasting errors. These reserves can provide extra power or take away excess power when needed. The Dynamic Reserve Setting project with the Smith Institute utilises machine learning to set reserve levels dynamically, a day ahead. This will allow the control room to more accurately forecast the levels of back-up power needed to maintain grid stability, while reducing operating costs and carbon emissions. 

Reducing consumer costs and achieving net zero

Our initial results show that by using the Dynamic Reserve Setting tool we can acquire, on average, 300MW less per reserve settlement period. By closely matching power reserves to the power requirements of the grid, the Dynamic Reserve Setting project can reduce balancing costs by avoiding repositioning generators when not needed to deliver value for money for consumers. Reducing our power reserves means we will be less reliant on fossil-fuelled generation to provide headroom, and we can better manage the need to take power off the grid without restricting renewable energy generation. This will help us achieve our ambition of delivering periods of 100% zero carbon operations by 2025, leading to a zero-carbon electricity system by 2035. 

Recognition from industry 

Our Dynamic Reserve Setting project has been shortlisted in three award categories at the DataIQ Awards ahead of the awards ceremony in September. We’ve been shortlisted for:

  • Most innovative use of AI (Artificial Intelligence)
  • AI-enabled data solution of the year (vendor-side)
  • Best data-driven process (client side)

View the full awards shortlists here.

What’s next for Dynamic Reserve Setting?

Phase one of the project delivered a proof of concept and phase two will look at improving on the initial designs and ensuring they work in reality.

Before the Dynamic Reserve Setting is transitioned into our daily activities, the demonstration tool will run in parallel with our existing reserve power recommendation tool for up to a year to ensure it is accurate. 

Working alongside the Smith Institute, we’ve shaped our previous model, so it can be used with features such as a review dashboard, and new functionalities are being considered for future implementation, using inputs from the Solar PV Nowcasting innovation project with Open Climate Fix, to further reduce reserve holdings from an improved solar forecast.

Want to learn more?

You can find out more about our Dynamic Reserve Setting project on the Energy Networks Association website here. View this page to view updates on the project milestones, a project summary and our lessons learnt. Visit the Smith Institute website here.

You can also find out more about Dynamic Reserve Setting, and our other innovation projects, in our Innovation Annual Summary.

Get in touch

Visit our website or contact us at [email protected] to learn more about our innovation projects and priorities.