Standard content for Members only
To continue reading this article, please login to your Utility Week account, Start 14 day trial or Become a member.
If your organisation already has a corporate membership and you haven’t activated it simply follow the register link below. Check here.
As more details emerge on the causes of last Friday’s power cut, the former head of engineering at National Grid argues the blackout should prompt a fundamental rethink of the strategy for dealing with outages and the growing volatility of the electricity system.
Last Friday shortly before 5pm the Little Barford gas power station Bedfordshire unexpectedly cut out following a technical issue. Failures like this are not out of the ordinary.
As per usual, the electricity system operator (ESO) at National Grid responded by activating its frequency response services to contain the initial disturbance. Within a few minutes, the frequency had bottomed out at 49.2Hz and had begun returning to the desired 50Hz level.
However, the rebound was short lived. Soon afterwards around two thirds of the Hornsea offshore windfarm went offline. Taken together, the two events removed more than 1.4GW of generation from the power grid.
With no further reserves to call upon, the ESO could do little but watch as the frequency plunged to just 48.8Hz, tripping a number of smaller plants on the way.
It was at this point that the Low Frequency Demand Disconnection mechanism kicked in, automatically shedding demand to keep the frequency from falling further and prevent a complete blackout. Across the UK around a million customers had their power supplies cut off.
Grid frequency during blackout
Source: Aurora Energy Research
They weren’t without electricity for long. Supplies were fully restored by all six of the distribution networks operators (DNOs) within 45 minutes.
But this was of little comfort to the thousands of people stuck at railway stations and on trains, many of whom were left stranded for far longer. Coming during the Friday rush hour as commuters headed home for the weekend, the timing of the blackout was hardly fortunate.
National Grid quickly found itself the subject of intense scrutiny. Newly appointed business and energy secretary Andrea Leadsom said the government would commission the Energy Emergencies Executive Committee to conduct an investigation. Ofgem requested an “urgent interim report” from the ESO by the following Friday (16 August) and a “final, detailed” report by 6 September.
The incident has placed a bright spotlight on the growing volatility of the electricity system. As renewables increase their share of the generation mix, the inertia of the power grid – its resistance to sharp changes in frequency – is declining.
Inertia is provided by synchronous generators that store energy in large rotating masses such as gas turbines spinning in harmony with the frequency of the electricity system. Most renewable generators are asynchronous and so cannot provide inertia.
Questions have been raised about whether low inertia was a contributing factor in Friday’s power cut, in particular whether the initial drop in frequency triggered the loss of mains protections at Hornsea.
The operator, Orsted, initially declined to comment on the cause of the outage, but has since confirmed that windfarm, like Little Barford, suffered a technical failure.
Speaking on Radio 4’s Today programme, National Grid chief executive John Pettigrew said: “The events of Friday – with two large generators falling off the network simultaneously – a very rare and unique event – were not related to renewables, or at least there’s no evidence at the moment to suggest it was.”
Nevertheless, there is still the question of whether low inertia contributed towards the ESO’s inability to contain the outages.
According a media briefing published by Aurora Energy Research, at the time of the failures nearly half of the generation on the system was asynchronous.
“There was an event the previous day where a different plant tripped but there wasn’t the same deviation in frequency and that’s most likely due the fact that there was much lower wind penetration,” says Aurora associate Harry Sturgess.
Given that periods of low inertia will only become more common over time, Aurora asked whether perhaps the ESO should have had more frequency response available.
Synchronous versus asynchronous generation
Source: Aurora Energy Research
“Being prepared for every possible eventuality may be expensive,” the briefing explained, “but we have seen that even short outages cause high levels of disruption and associated cost if key infrastructure such as airports, hospitals and railways are taken out.”
The ESO currently spends around £170 million annually on frequency response services. Aurora said it could double this spend and add just £2 to the average household energy bill.
Sturgess tells Utility Week: “There’s a huge amount spent on the capacity market and that provides you with element of security of supply that is meeting peak demand in winter.
“But then there’s the other element of security of supply that is managing the system frequency response and being able to respond to outages. The amount spent on that if you look at frequency response alone is significantly lower.”
Steve Shine, executive chairman of the renewable developer Anesco, was far less equivocal about the lessons that should be taken from the blackout, saying the blame “lies squarely” with the ESO for failing to secure sufficient frequency response.
But Cornwall Insights senior consultant Thomas Edwards questions whether the additional security would be worth the extra cost: “A lot of the problems that occurred were due to the inability of the rail network to cope with a loss of power. Everything was back online within 40 minutes.
“If you were to cover the whole of the largest infeed loss requirement with a sub-one-second frequency response service; if you were to cost that based on the recent enhanced frequency response tenders, you been looking at a probably half a billion pounds.”
He notes that the government’s own security of supply standard allows for three hours of supply interruptions each year – around four times the length of Friday’s power cut – and asks whether the money would be better spent improving the resilience of other critical infrastructure like railways.
John Scott, the director of the consultancy Chiltern Power, says whatever conclusion the ESO reaches must be based on hard facts and analysis: “It might sound like a good thing to do but unless you’re tackling the root cause of the problem it’s money that’s not well spent.”
Like Edwards, Scott says there should be a thorough examination of why the ESO’s load-shedding caused so much disruption: “Supply can go off at any time because of a storm or something like that. It’s in a smaller area of course but it’s often off for a lot longer. Whereas on Friday it happened over a widespread area and it affected an awful lot of people.
“But why, given that power was restored in about an hour to the distribution companies, did it take so long for their customers to bring their services back?
“And it would suggest that there are some really important lessons for the railway industry and other critical sectors to look into.”
Some commentators have asked why the railways were among those to lose their power supplies.
But Scott says, whilst issues like these are worth exploring, they are merely tactical in nature. He says last Friday’s blackout demonstrates the need for a more fundamental re-think of the ESO’s strategy for maintaining the stability of the power grid.
He says the Low Frequency Demand Disconnection mechanism has remained largely unchanged since its creation in the 1960s and is now showing its age: “We’re moving to a world with much more distributed energy resources and demand flexibility, which we need anyway for managing more renewable energy on the system.
“Shouldn’t we begin to think about whether that flexibility could be part of the defence plan?”
Data-driven intelligent system
He continues: “So, for example, we’ve got a lot of batteries starting to appear. We’re going to have millions of electric vehicles, which when they’re plugged in will be like a mega-battery.
“How about a data-driven intelligent system that monitors which batteries are available and have got power available in them and that at that moment of emergency, instead of shedding demand, says switch on your battery, and we’ll pay you for it?”
Scott says there should also be more work to explore whether asynchronous generation could be used to provide so-called synthetic inertia: “We’re losing inertia as big central power stations close. Those stations are being replaced by wind farms and solar arrays which don’t have mechanical inertia that is offered into the system.
“But couldn’t the power electronics create a form of performance that looks to the system like inertia?”
Synthetic inertia is essentially a form of fasting-acting frequency response. In the case of a wind farm, a drop in frequency prompts the turbines to convert more the blades’ kinetic energy into electricity and feed it into the grid.
“Synthetic inertia is simply very sophisticated programming of power electronics controls,” he explains. “With the right signals, they can be made to provide dampening to the system”.
As Scott notes, this is not a new proposal. Indeed, Hydro-Québec, a public utility in Canada, has since 2005 required all new wind turbines to be capable of boosting their power output by up to 6 per cent of their capacity in response to a drop large drop in frequency.
The first turbines with this capability were installed in 2011. By 2016, they made up two thirds of all wind turbines in the region.
Scott says these suggestions highlight a wider issue over the lack of coordination across the energy system. Both would require collaboration between the ESO and DNOs as well as numerous other players.
“At the moment, it’s nobody’s job to provide coordination across all those parties,” he laments.
What is needed, he says, is a system architect. He says the energy industry needs to overcome its fear that such a body might eventually morph into something like the Central Electricity Generating Board from the days before privatisation.
“The system architect is not a central planner, because they don’t own assets,” he remarks. “Their responsibility is to coordinate to make sure the total system works.”
Scott says the same issue is emerging as the ESO seeks to enable distributed generators to provide the black start services used to reboot the power grid following a complete shutdown.
Regardless, Scott says the industry should seize on the power cut as an opportunity for reflection: “Nobody’s asking anyone to jump into a solution, but let’s jump into having a discussion about the question with some fresh thinking.”
Please login or Register to leave a comment.