<aside> 💡 This is the methodology for the app deployed at https://wind-curtailment-app-ahq7fucdyq-lz.a.run.app/

This is a side project done on weekends and evenings, so there are a few rough edges. If you spot something that doesn’t look right to you, we’d really appreciate you getting in touch.

</aside>

Core analysis behind the app

• Data comes from the Elexon Balancing Services API, via the helpful ElexonAPI package (source)
• We combine data from
  • FPN (Final Physical Notifications)
    • We use this to work out what wind assets could be generating, if they were not instructed otherwise by the grid
  • BOAL (Bid Offer Acceptance Levels)
    • We use these to identify which assets have been turned up or down by the grid, and to what level. The difference between the FPN and the BOAL is the curtailment or redespatching level
  • BOD (Bids and Offer Data)
    • We use these to understand how much the grid paid for turbines to turn off
  • SIP (System Imbalance Prices, or System Buy Prices)
    • We use these to estimate the cost of turning up alternative generation to offset curtailed wind. We have used £200 per MWh in 2022 and in other years we have used £100 per MWh.

Secondary analysis

In the blogpost we also draw upon:

• The Renewable Energy Planning Database Quarterly Extract to specify the locations of wind farms. We then manually mapped these to BM units, since there was no clear algorithmic mapping. Here’s the file in case it’s useful to you:

  bm_units_with_locations.xlsx

• Some total generation figures from the Elexon Insights API (this one)

Code

We would like to make the code open source but we need to clean up some API keys first, and write some documentation. If you have particular interest in this, please let us know.

Update: The code is now public - https://github.com/hackcollective/wind-curtailment