GB power imbalance volatility: a deep-dive

The GB system imbalance is volatile. On a nearly daily basis, it to repeatedly flips between excess generation and shortfall generation over the course of a single day, but what drives this?

Our experts work with our clients to help them understand the power markets all the time. Recently, a client asked us to provide a detailed overview and understanding of what drives this volatility in the Net Imbalance Volume (NIV). To do this, our analysts used our Enact power market analytics platform, and deep dived into a fairly-regular day – 9 July 2023.?

While nothing unusual occurred on this particularly normal day, we thought it would be useful to do a deep-dive to demonstrate how even routine events can drive volatility intraday changes.

Here's our summary of this deep-dive, written by Tim Sparks , into how remarkable an unremarkable day on the GB system can be:

?? Background

The GB system often flips repeatedly between long (generation is greater than demand) and short (demand is greater than generation) over the course of a day.

This volatility in the day is more likely to happen when the market does a good job of dispatching generation and leaves the system quite close to balanced, meaning that it only takes small changes to send the system imbalance one way or the other.

Alternatively, as with 9 July 2023, it could be an eventful day on the system where there are multiple events which significantly impact the balance of supply and demand. For example, demand or renewable output being significantly less than forecast or there is an unplanned outage which throws the imbalance significantly out.

National Grid ESO produce a forecast of imbalance (known as IMBALNGC) which reflects how well the generation dispatched through the wholesale market (known as INDGEN or Indicated Generation) lines up with ESO's demand forecast (known as TSDF or Transmission System Demand Forecast). NIV Outturn reflects the actual system imbalance based on the actions National Grid had to take to bring the system into balance. This essentially means that in the Balancing Mechanism once the volume of bids and offers has been netted off, we can work out whether the system was short or long and by how much.?

This forecast is understandably quite important for those trading and operating in the market, which is why we at LCP Delta provide our own enhanced NIV forecast that can give our clients a competitive advantage.

Drilling into 09/07/2023

???? Start of day - Wholesale Market

This chart shows the ESO’s imbalance forecast in comparison to the NIV outturn. On this day the wholesale market was forecasted to position the system long based on the latest available forecasts at the time of gate closure (when all BM Unit assets have to declare what power they sold or bought). This is demonstrated by the ESO’s imbalance forecast was mainly positive in this chart in pink.

However as we see in the chart, the actual outturn on this day frequently flipped between long and short; in stark contrast to that forecast by ESO. This was due to several factors, of which we're going to focus on the morning:

?? 00:00 to 02:00 - Wind outturns below forecast

Looking at the imbalance chart, we can see that the system was short in the first two periods (half-hours) of the day. Now looking at the wind chart, we can see that wind generation (transmission wind outturn) was ~1GW below forecast at this time, partly explaining why the system was actually short.

After this, wind generation then mostly converged back in line with forecast, helping to reduce system imbalance deviations.

???? 06:30 - Unplanned gas (CCGT) outage

Imbalance volumes were then relatively small until 06:30 when a CCGT (gas) plant had an issue which meant they couldn’t generate to their Physical Notification (PN, i.e. what they sold in the wholesale market).

As seen in the chart above, this produced over 200MW of unexpected unavailability. This pushed the system short as ESO had to dispatch generation to replace this CCGT generation error.

? Post 06:30 - Demand ramps

We then see that demand is ramping up towards the morning peak and this helps the system return towards balance. This was despite continued wind over-delivery and a reduction in the CCGT generation error (as the plant returned to expected performance).

?? 11:30 - Demand drops as embedded solar over-performs

From around 11:30, we can see in the above demand chart that demand starts to fall well below forecast. Looking at the solar chart, we can see higher than expected levels of embedded solar generation. Embedded generation feeds into demand so causes this deviation.

This sees the system go significantly long over this period of peak solar output (longer than suggested by the imbalance forecast from ESO). Although, this was offset in part by some further CCGT generation error (see the gas outage chart above).

Post midday - continuing errors

From 14:30, demand is recovering back towards the forecast level, while the CCGT (gas) generation error persists and wind has fallen back below forecast. This sees the system imbalance flip from long to short in the period starting at 15:00.

The CCGT generation error reduces and wind output then converges back towards forecast as the system goes long again in the period starting 17:00.

The NIV outturn roughly mirrors the ESO's forecast imbalance in the run up to the evening demand peak, until wind and more CCGT generation error (due to more unplanned outages) push the system short from 20:00 through until the end of the day.

?? Summary

The energy markets are inherently complex and unpredictable. As we've now seen, even the volatility seen on a seemingly unremarkable day such as the 9 July 2023 in the GB system, was driven by a blend of different factors.

These factors highlight the constant changing nature of the energy market and the challenges faced by energy operators in ensuring balance. They underline the importance of accurate forecasting tools helping traders to better position themselves in the market.

Let us know in the comments if there any other events you'd like us to drill into!

Want to learn more about Enact? Reach out to Rajiv Gogna or Shivam Malhotra , and explore for yourself here: LCP Enact.