Blackouts near miss in tighest day in GB electricity market since 2011

Yesterday saw a blackout near miss in what turned out to be the tightest day the GB electricity market has seen since 2011. Wind power was 2.5 GW through the evening peak, solar was (obviously) zero and there were significant interconnector outages leaving expected capacity at just 5.7 GW. Had just one large power station tripped this evening, demand control would have been a real prospect.

In this post I will walk through what happened, and how close we came to wide-scale loss of load (demand control means the National Energy System Operator, NESO, will disconnect parts of the country to avoid uncontrolled blackouts). Friday is also looking tight so it’s worth understanding what happened on 8 January, whether the situation could have been managed better, and just how close we came to demand control.

An unfolding challenge

Market forecasts eg the ones available on Amira Technologies’ dashboard (see below) had been signposting tightness for Wednesday 8 January and Friday 10 January. The snapshot below was taken at around 8am on the 8th, and shows this expected tightness – similar snapshots last week showed similar expectations for Wednesday and Friday. So the market saw some of this coming.

The next clue that things were tight was that NESO issued the first Electricity Market Notice of the winter at 8:36pm last night:

An ELECTRICITY MARGIN NOTICE has been issued by the National Energy System Operator to encourage market actions to increase System Margins. For the period: from 16:00 hrs to 19:00 hrs on Wednesday 08/01/2025 There is a reduced system margin. System margin shortfall 1700 MW The current contingency requirement is 1000 MW. 300 MW of generation is excluded from the available system margin due to system constraints.

Maximum Generation Service may be instructed. Trading Points, Control Points and Externally interconnected System Operators are requested to notify National Energy System Operator of any additional MW capacity. Suppliers please advise National Energy System Operator of any additional Demand Control available. The situation will be reviewed again by National Energy System Operator at 07:00 hours and an update issued.

This was updated at 6:59am today to read:

System margin shortfall 1700 MW. The current contingency requirement is 900 MW. 0 MW of generation is excluded from the available system margin due to system constraints

And again at 11:58am to:

System margin shortfall 1120 MW The current contingency requirement is 448 MW. 0 MW of generation is excluded from the available system margin due to system constraints.

Before being cancelled at 4:19pm, after the start of the period covered by the notice.

In other words, as the day went on curtailment was cut to zero and additional generation found to reduce the margin shortfall.

A Capacity Market Notice was issues at 12:01pm and cancelled half an hour later.

What is an Electricity Market Notice?

An Electricity Market Notice (“EMN”) is issued when the spare capacity does not cover the contingency NESO has decided is necessary. For example if the contingency requirement is 1000 MW but there is only 500 MW of spare capacity, there would be a shortfall of 500 MW.

“If we can see that our normal safety margin for operating the system is not as big as we’d like, and we can’t address it through the normal mechanisms, then we would consider issuing an Electricity Margin Notice… If our safety margins for operating the system are reduced, a Capacity Market Notices (CMN) might also be issued as an alert to providers in the capacity market,”
– NESO

While EMNs and CMNs are based on the same fundamental data including generator availability and demand forecasts, they are issued based on different thresholds and with different lead times. EMNs are issued by control room staff using operational and engineering judgements, based on their “experience, skill and knowledge of managing the electricity system”. CMNs are triggered automatically four hours ahead of real-time based on specific industry data about the system’s safety margin.

This means EMNs and CMNs can be issued and active at different times. It’s not unusual for a CMN to be triggered and for the control room staff to scratch their heads because they do not consider the system to be tight.

The chart illustrates when and why an EMN and CMN might be issued during the 24 hours up to delivery. The coloured line indicates the forecast buffer of surplus capacity above demand and the operating margin that NESO is required to hold to operate the system when it comes to delivery.

When an EMN is active, any generator that has technical availability to run must make itself available in the Balancing Mechanism (“BM”). Normally periods of EMNs coincide with high market prices, so there is a strong incentive for generators to run, but if they are concerned about reliability, the costs of trips would similarly be high which could make them reluctant to take the risk. There can be other reasons not to run, such as wanting to delay triggering maintenance (although again, when prices are high there would be a strong incentive to run that day and not run another day despite potentially being in the money). In any case, there are legitimate reasons that a generator may not plan to run despite being technically available. During an EMN, this choice is removed, but the generator is able to choose its Bids and Offers in the BM, and can select high prices if it has concerns about running.

During an EMN, NESO also asks suppliers about demand control availability. This is to help NESO to get a better picture of the demand that can be saved through demand control.

Demand control can become necessary under two different conditions – a sudden or short-duration electricity shortfall affecting a specific region, or the whole country, or a prolonged shortfall affecting a specific region or the whole country. The latter case is the one which could arise on days when the market is tight. The former usually occurs when something breaks or trips suddenly without notice.

If a prolonged shortfall is expected, NESO and the Distribution Network Operators (“DNOs”) would be directed by the government to implement rota disconnections using the process defined in the Electricity Supply Emergency Code. This means different regions can be taken off supply for periods of four hours at a time until the issue is resolved. This process ensures fairness, so that consumers are affected evenly, while protecting those who require special treatment. Once invoked, this process allows the government to direct NESO to implement rota disconnections for up to 90% of demand at 1 days’ notice.

In the context of an EMN, it’s unlikely that there would be a day’s notice at the start. Today, had there been a significant trip during the evening peak, such as the NSL interconnector (1.4 GW) then there could have been a sudden frequency event. If a smaller unit such as a 500 MW gas plant tripped, this may reduce the spare margin to such a degree that NESO decided to implement Demand Control to prevent the risk of a country-wide blackout. It would use the data provided by suppliers and DNOs, as well as its view on grid constraints and other physical considerations, when deciding which regions should be disconnected.

A High Risk of Demand Control (“HRDR”) notice or Demand Control Imminent (“DCI”) notification would be issued by NESO once it became clear that demand control was becoming likely.

What happened on 8 January 2025?

On 5 January, NESO published its peak demand forecast for 8 January at 43.267 GW. Generator availability was expected to be 47.425 GW, and interconnector availability 6.674 GW. The operational planning margin requirement was 3.235 GW, meaning the surplus was 0.923 GW.

At 8am on 8 January, peak demand for the evening peak forecast by Amira to be 48.6 GW – this is intended as a true GB electricity demand forecast accounting for the effects of embedded generation in the distributions networks. At the time, the NESO transmission system demand forecast was 46.777 GW (at 17:00). The Initial Transmission System Demand Outturn data from BMRS indicated that actual transmission system demand was 46.825 GW (at 17:30) ie the forecast was too low.

The chart below indicates rolling system demand for the day – the maximum was 47.088 GW. Rolling system demand includes all demand met from the transmission system including power station auxiliary demand, pumping demand and interconnector export demand. The data is provided as 5 minute average MW values, half hourly averages in MW and daily MWh totals. As this is related to the breakdown by fuel type of total positive generation to meet all transmission system demand, any fuel type categories with negative values will be capped to zero.

NESO’s demand forecasts can be quite inaccurate. It publishes demand forecast performance analysis (not very often) – half-hourly forecast versus out-turn data from April 2021 and November 2024 indicate the forecast error can be from 0 MW to 4,686 MW with an average of 609 MW. This magnitude of error could be critical on a day like today.

The generation availability (data from Amira Technologies taken from BMRS) was 44.905 GW. Adding the expected wind generation of 2.5 GW, that gave 47.405 GW. The out-turn after adjustments in the BM was 46.628 GW which includes an additional 700 MW from the Viking interconnector which was able to return a bipole that was intended to be offline for maintenance.

However, there had been a bit of spare CCGT capacity – the availability was 26.007 GW, the out-turn before BM actions was 24.368 GW and after BM actions the out-turn was 23.230 GW. Some CCGTs, notably Rye House, were offered at such high price in the BM they were only partially accepted. Rye House ran at its Stable Export Limit rather than its Maximum Export Limit with the remaining 300 MW held in reserve, while Dinorwig pumped hydro station was taken out of reserve and run instead since it was available at lower cost.

So to summarise, the total available supply at 5:30pm – generation plus interconnectors – was 47.405 GW while the peak demand at this time was 46.825 GW. This means that the actual spare margin on the system during the peak was just 580 MW! Even a relatively small power station trip would have caused an actual shortage and triggered blackouts. Had Viking not returned to full service it would not have been possible to meet peak demand.

NESO did have one other tool available to possibly increase generation, as noted in the EMN: “Maximum Generation Service may be instructed.” Known as Max Gen, this is an instruction to power stations to run at their maximum output irrespective of their Maximum Export Limit (“MEL”). Normally power stations could get c10 MW more from a turbine than its MEL and if every power station did this, the output could be meaningful.

Unfortunately, when this has been activated in the past, power stations tended to trip off, because once they are pushed to their operational limits, some are unable to cope and break down. This means that activating Max Gen comes with significant risk, and these days is considered to not be worth it –  a trip of a larger unit could wipe out or exceed the gains from all the other generators, possibly making things worse than if it had not been activated. I understand demand control would be implemented without attempting Max Gen these days.

Did the Winter Outlook predict this?

The answer is a resounding “No!”. The Winter Outlook published in October proudly announced that the capacity margin was higher than last year (and for this reason the Demand Flexibility Service (“DFS”) was designated as a normal margin management tool and not part of an emergency contingency – the DFS was not activated on 8 January). The Winter Outlook predicted that weather-corrected peak transmission system demand would be 44.4 GW, down from 44.9 GW last winter. Actual peak demand on 8 January was 46.8 GW – 2.4 GW higher than the peak demand expectation for the winter. This is a huge error.

NESO assumed that 6.6 GW (de-rated) net imports woud be available via interconnectors at times of tighter margins – in fact only 6.26 GW of interconnector imports were secured on 8 January.

The error on the de-rated margin was even higher – the capacity margin was 5.2 GW but the actual spare margin was only 0.58 GW – that’s an order of magnitude off! The surplus daily credible lower bound for the winter was 1.77 GW – more than three times higher than the actual surplus!

A worrying lack of transparency

Unfortunately, NESO is not transparent about just how close the market came to disaster today. While it is possible to find all of these data from public sources as I have (Amira gets its data from BMRS and NESO, and I used it to save time, wanting to publish this blog sooner rather than later), but the official line from NESO seems to be “it’s all fine, nothing to see here!”

A NESO spokesperson told The Telegraph:

“Both of these notices are part of Neso’s operational toolkit and are routine tools used most winters. These notices are to inform the energy industry of the need for additional electricity supplies to maintain the operational reserve Neso holds whilst operating the national electricity network. These notices do not mean that electricity supplies are at risk or that there is not enough electricity to meet demand.”

This very much under-states the near miss on the day.

There are other problems, particularly relating to interconnectors. Some of the interconnectors can re-trade within day, and in the worst case, the control room may only have 55 minutes’ notice of the trades. If such a trade reversed the flow, an interconnector could reduce the level of imports or even switch from imports to exports at extremely short notice. NESO would have limited ability to react – while it can and does buy up interconnector capacity to secure imports, this trading takes time to execute – and can cost a lot of money. If there is only 55 minutes’ notice, the only real option would be to ask the interconnector operator to implement an Emergency Action to secure the imports, but this can be declined and NESO cannot force it.

Interconnector owners do not control the flows under normal operations – they make capacity available and traders in the market will buy it, flowing power one way or another. This is why, as I have noted in the past, the inclusion of interconnectors in the Capacity Market is inappropriate and provides a false sense of security. Under Capacity Market rules, interconnectors must be available in a system stress event, but being available means electricity can flow, it does not mean electricity will flow to GB – the interconnector will meet the availability criteria if it is exporting.

There are 38 registered trading parties on the interconnectors. Any one of them could enter a late trade when the market is tight – they may do this for portfolio reasons, or they may do it maliciously, or to make a point about the market operation. NESO has no control over this.

Another issue is cost. The average daily balancing cost is £2.3 million, but on 8 January more than £21 million was spent on balancing actions by NESO, according to LCP’s Energy Current dashboard (which unfortunately does not show historic data on the free version and I don’t have time to dig this out on BMRS to demonstrate it here). Essentially the balancing costs for the day were close to 10 times the normal daily costs. However, the costs of blackouts would be much higher. Balancing Mechanism prices reached £5,500 /MWh and there were several settlement periods with a cashout price of £2,900 /MWh.
.

On 8 January the GB power market came within a whisker of blackouts. NESO used almost every last MW available with just 580 MW of cushion – only two thirds of the contingency that should be held, and no-where near the single largest infeed loss which is supposed to be protected. And securing that minimal spare margin cost more than £21 million. This should be a real wake-up call about the dangers of relying on weather-based generation, but because NESO tells everyone things are fine even market participants may not realise just how close we came to demand control or blackouts.

NESO has a duty of transparency, so it is disappointing that it relies on people taking the time to dig into the market data to understand the situation, rather than being explicit about it, although I do understand there will be a post mortem in the Operational Transparency Forum call next week. To join the call, register on the link – it takes place at 11am every Wednesday. It will be interesting to see what is said and whether NESO owns up to the near miss on 8 January.

And now we face the same again on 10 January which is also forecast to be tight…