Here’s what we know about the events beginning on Friday afternoon at around 4:52pm.
Flexitricity was having a quiet afternoon, with most of our customers getting on with the day job, and just a few megawatts of Short Term Operating Reserve running at the time.
We were also charging some battery capacity under Balancing Mechanism instruction.
National electricity demand was a modest 28GW when, for as-yet-unannounced and probably mundane reasons, Little Barford power station tripped.
It had been generating 664MW.
Little Barford is a combined-cycle gas turbine (CCGT) of nineties vintage, snuggled next to the main east coast train line between London and Edinburgh. It’s owned by RWE.
At around the same time, Orsted’s brand new Hornsea offshore wind farm also tripped.
Orsted’s data housekeeping is perhaps a little less neat than RWE’s, so it’s tricky to ascertain exactly how much was lost (apparently at least 800MW), or whether Little Barford or Hornsea went first.
But the combined effect was certainly larger than the largest single-infeed loss that National Grid would have been planning for – presumably the French interconnector, which was importing around 1000MW at the time.
At other times, they might have been watching Sizewell B, which can hit 1250MW, but our most modern nuke was running at reduced power at the time.
Mains frequency plummeted to around 49.1Hz, wobbled, then took the final drop to the red line of 48.8Hz. At this level, there is an obligation on Distribution Network
Operators to shed around 5% of demand, automatically and quickly. They do this with low-frequency relays around their networks, all set up in advance. This happened as planned.
Demand fell by about the right amount, and frequency stopped falling. As Duncan Burt, National Grid’s Director of Operations, put it on the news on Saturday morning, “the system worked”.
From that point of view, it worked very well indeed. Frequency didn’t take long to recover – just long enough for National Grid to flip our battery charging into discharging, helping to fill in the power gap – but it rose so fast from there that they had to flip us back into charging mode. It’s worth remembering that excessively high frequency can be just as bad as low. Both can lead to nationwide blackouts.
This is what Duncan didn’t say. If the 5% demand cut hadn’t been enough to stop frequency falling, more demand would be cut at different frequency levels until the power stations themselves gave up trying.
That would have led quickly to a nationwide blackout, and the one circumstance that really keeps National Grid awake at night: a black start.
About a million people were affected by the power cuts that did occur, but their best chance of getting power back quickly was to avoid losing it nationally.
It’s relatively easy to reconnect customers to a stable system; it’s a whole lot harder to restart a system where absolutely everything is off.
There will certainly be investigations, and they will be detailed. One thing which will receive a lot of attention will be the report dealing with strikingly similar events on 27 May 2008, when frequency fell to exactly the same level of 48.8Hz, following the near-simultaneous trip of two power stations.
On that day, it was coal (Longannet) and nuclear (Sizewell B), rather than gas and wind.
But both days saw a noticeable wobble in frequency as it headed south. The 2008 conclusion was that falling frequency had destabilised a number of distribution-connected wind farms.
This led, one way or another, to the current project to upgrade protection settings at distributed renewable generators.
If the same thing happened this time, there will be questions about the speed of that project, and whether the networks are doing enough to engage with their embedded generator customers and get this done in a collaborative way.
But it’s likely that much of the investigation will focus on something else: affected customers.
Industry insiders need to remind themselves often that customers are real, that electricity is very important to them, and that failing to supply it can be really, really bad.
Two subjects will get special attention. The first is Ipswich Hospital, where a back-up generator reportedly failed, although battery-based uninterruptible power supplies successfully supplied critical areas.
The woeful reliability of Britain’s standby generator fleet is something that Flexitricity has blogged about in the past, and as always, our solution is a proper test-and-exercise regime tailored to national and local need for electricity.
This is something that recently became quite difficult: DEFRA and the Environment Agency got their wires quite seriously crossed when implementing the European Medium Combustion Plant Directive, making good testing much harder.
The second is trains. Since the distribution networks have a free choice as to which circuits to trip at the 48.8Hz trigger level, did they really have to do it to the trains?
The May 2008 event caused a similar level of power cuts, but had far less impact on rail transport. So was the loss of traction power a direct result of the low frequency trip, or an unexpected secondary consequence?
Perhaps more importantly, where was the resilience in the signalling system? This is a much smaller load than traction, and should be amenable to high-specification backup at modest cost.
I would be interested to know if anyone was stuck just south of St Neots on the east coast main line.
If they were, they’d have been looking right at one of the two main sources of Friday’s blackout blues.
But power stations like Little Barford sit down all the time.
It’s prosaic; it’s planned for; and filling gaps like that is what Flexitricity does every day.
Understanding why this particular event got so bad so quickly will be the subject of much detailed technical analysis.
The solution won’t be to build lots more gas power stations. It wasn’t for want of fossil-burners that so many people took so long to get home on Friday night.
On the subject of getting around, I found myself in the BBC’s Inverness studio at 10pm that night talking about the blackouts.
I got there in a fully electric campervan, which I’d charged that day using 100% wind energy. And that worked absolutely fine.