August 12, 2019

expert reaction to the major UK power failure

On Friday 9th August the UK experienced major electrical power failures effecting transport and infrastructure across the country.  

Prof Keith Bell, Electronic and Electrical Engineering, University of Strathclyde, said:

“On the global scale of electricity system disturbances, e.g. in Jakarta and Western Java at the weekend and in Argentina in June where practically the whole country was blacked out, this has been a relatively small event.  However, even though National Grid has stated that “By 6.30pm, all demand was restored by the distribution network operators,” this was still massively disruptive for lots of people largely, it seems, due to time taken to restore both train and signalling supplies to the railways.  Then, many trains will be in the  ‘wrong places’ relative to the timetable.

“A worst case outcome of any disturbance for an electricity system operator is that the whole system goes down.  Recovery is then massively challenging, not just for users of electricity such as the railway companies but also the power system operators.  This event is perhaps a reminder that, even though we have never suffered a whole system blackout in Britain and many of our system design and operation procedures have stood the test of time, restoration plans always need to be kept under review.

“The nature of the system continues to change with, quite rightly in view of our emissions reduction commitments, more low carbon sources of power being used.  This means that the normal operating procedures and the codes and standards that govern the system also need to be kept under regular review, especially in light of our increasing dependency on electricity.”

Prof Tim Green, Co-director of the Energy Futures Laboratory, Imperial College London, said:

“This event does not appear to be due to wind generation reducing owing to reduced wind speed.  If that were the case there’d be reduction across many wind farms in same area.

“First generator to disconnect was a gas fired plant at Little Barford at 16:58.  Two minutes later Hornsea Offshore wind farm seems to have disconnected.  This would seem to be a technical failure/error.  Might be linked to disturbance caused by first generator failing; might not.  Will need to wait for National Grid’s full technical investigation to get to bottom of that.

“There was a lot of wind power today and consequently less gas used.  A system with little gas plant running does need careful management because when generators fail the frequency responds more quickly.”

Prof Jon Gibbins, Professor of CCS at the University of Sheffield, and Director of UK CCS Research Centre, said:

“Looking at 5 minute data from Gridwatch https://www.gridwatch.templar.co.uk, solar power was decreasing as the sun went down at the time of the problems on the UK Grid, apparently in fairly large, abrupt steps in the amount of solar power actually delivered coinciding with half hour bidding periods.  Possibly something didn’t come on reliably to replace the lost solar power at one of these steps, frequency dropped and there was not enough reliable and responsive power sources already up and running to kick back in time before load had to be shed to stop grid frequency dropping too far.  And perhaps the low frequency also caused some other generation sources to trip.  But the Gridwatch data resolution (5 minute intervals) is not enough to show what happened in detail.  That would need data at intervals of 5 seconds or less as things can happen very quickly in the electricity supply grid.”

Prof Keith Bell, Electronic and Electrical Engineering, University of Strathclyde, said:

“Although there are some questions that could be asked about the event and potentially lessons to learn, there’s no indication that this event has anything to do with lack of investment in major infrastructure.

“Also, there is no indication that this is anything to do with the characteristics of wind as a source of electrical energy. Even if, as some reports are saying, one of the two ‘loss of infeed’ events was at Hornsea wind farm, it seems likely, based on the frequency trace that colleagues at Strathclyde have available from the lab, that it was due to something like a fault on the connection to the main grid rather than due to any changes in wind speed.”

Prof Tim Green, Co-director of the Energy Futures Laboratory, Imperial College London, said:

“National Grid’s statement that it had “issues with two generators” would explain what we’ve seen.  If a generator has to unexpectedly shut down the frequency of our AC system starts to fall.  It’s normally close to 50 Hz but does vary.  National Grid has to respond quickly to such a fall with generators that are held in reserve, notably those contracted to provide “frequency response”.  But if a second generator shuts down soon afterwards then the further drop in frequency might have to be counter-acted by “load shedding” which is the term for disconnecting some customer to keep the rest of this system running.

“Upside Energy tweeted a graph of frequency that they recorded that shows a deep dip to about 48.8 Hz.  We’ve seen that once before on 27^th May 2008 following a failure at Longannet coal power station and 3 minutes later Sizewell B nuclear station.  On that occasion 500,000 people were affected by automatic disconnection.

“Upside Energy are one of the new companies in the energy field.  They can sell services to National Grid so that for a price, they will voluntarily disconnected some electricity users who can start up their own stand-by generators or run off batteries or perhaps shut down part of their industrial process.  This voluntary reduction saves others from involuntary disconnection.”

*Prof Green’s comment was amended to correct ‘We’ve seen that once before on 27^th May 2019…’ to ‘We’ve seen that once before on 27^th May 2008…‘ (11:16, 16/08/19)

Prof Keith Bell, Electronic and Electrical Engineering, University of Strathclyde, said:

“The news from National Grid seem to confirm my hunch. My next question would be: were these two trips quite random and independent of each other? If so, we were just unlucky that a second trip occurred so soon after the first one and before there had been a chance to replenish the ‘frequency containment reserve’ that had been used after the first event. (It seems to have been a very similar event to one in Britain in May 2008. However, this time, we were kind of unlucky in that (a) it was a deeper drop in frequency so more demand was tripped and (b) it happened during rush hour).

“It’s fair to say that, from a transmission system point of view (National Grid’s responsibility), things have now been returned to normal, but the relays tripped on the distribution networks. One likes to think that, with the green light from National Grid that things are ok at the transmission level, they can restore supplies pretty quickly. But how good have they been at managing the flow of information to ordinary people and, in particular, managers of facilities such as railway power supplies? (Note that there are, I think, 8 different DNOs across Britain. Some may have done a really good job. Others perhaps less so. I’m sure Ofgem will be trying to find out in the forthcoming investigation!)”

Prof Keith Bell, Electronic and Electrical Engineering, University of Strathclyde, said:

“An AC power system must be operated with a certain band of frequency, in our case near to 50 Hz. This requires that, moment by moment, generation and demand are matched. Automatic controls on many generators adjust their outputs automatically to fine tune the balance and National Grid ESO brings on extra generation or reduces generation through the day to get the broader balance.

“It looks like a major source of power onto the electricity system has tripped, e.g. a large generator or one of the interconnectors importing power from the continent. (I haven’t heard quite what or why but it’s not totally uncommon).

“This causes system frequency to drop below 50 Hz. National Grid ESO schedules “frequency response” – also known as ‘frequency containment reserve’ – on the system to be ready to respond to such drops and automatically inject more power onto the system to correct it.

“Frequency response seems to have kicked in and contained the fall to within acceptable limits and started to restore the frequency back towards 50 Hz when something else appears to have happened knocking it back down.

“There are automatic emergency ‘defence measures’ in place on the system that, in the event of system frequency dropping very low – a sign that at that moment, generation is significantly lower than demand – automatically trip some demand to restore the balance.

“These ‘low frequency demand disconnection’ (LFDD) relays are intended to trip only just enough demand, but in the event that, for whatever reason, they don’t succeed in restoring system frequency, further blocks of demand are tripped at additional, lower frequency thresholds, the purpose being that, even if some demand is disconnected, the system as a whole is saved – the worst case is a collapse of the entire system, such as happened in Argentina in early June, and this can happen, when a point of no return is reached, in seconds. (It appears that in that case in Argentina, among other things, their equivalent of LFDD didn’t work entirely as intended).

“System frequency appears to have dropped just low enough to trigger the first threshold of LFDD, at 48.8 Hz.

“The LFDD relays that automatically disconnect demand are installed and maintained by the Distribution Networks Operators (DNOs) through an industry agreement. As far as I know, the DNOs decide which circuits to install them on. Each relay has the effect of disconnecting a chunk of demand, but these relays are spread across the country, which is why the effects have been seen across such a large area while the system as a whole has continued to operate.

“National Grid will then have been trying to bring on extra generation to restore system frequency and also reserves in readiness for any other event that might happen. The circuits that were tripped by LFDD should then be reconnected by the DNOs.

“The good news is that the system as a whole has continued to operate. The defence measures can be seen to have succeeded in their basic aim. However, there are some questions, not least: what caused the second fall in frequency? It seems to me that there might also be some questions to the DNOs about where their LFDD relays are located.”

Declared interests

Prof Keith Bell: “I hold the Scottish Power Chair in Smart Grids at the University of Strathclyde. (Scottish Power sponsors the Chair, not me. I speak as an independent academic. I have had various research projects over the years with the electricity companies, but only on the condition that results are published and, from my point of view, with the aim helping to improve engineering practice. I have also done work with Government and Ofgem). See ://www.strath.ac.uk/staff/bellkeithprof/ (As as Chartered Engineer I have to uphold certain principles of honesty, etc.).”

Prof Tim Green: “National Grid do support some research at Imperial that I have a role in. I am free to speak though.”

None others received.