837: The 9 August power cut – what happened and what next?

Today’s entry analyses the serious power cut that occurred on 9 August.

In the early evening of 9 August, my birthday – you will be relieved to know that it did not interrupt the festivities – there was a power cut affecting more than 1 million people, a hospital and an airport and which resulted in 1,500 trains being cancelled. What happened and what lessons can be learned?

The National Grid Electricity System Operator (ESO) has published an interim report at the behest of the regulator Ofgem and must publish a final one by 6 September. I’ve had a read and set out the main points below.

Three things happened in quick succession. First, three bolts of lightning struck the electric line between Eaton Socon in Cambridgeshire and Wymondley in Hertfordshire. This caused about 500MW-worth of small electricity generators to be briefly disconnected from the rest of the grid. Secondly, part of the Hornsea One offshore wind farm, 24 miles off the Yorkshire coast, which is in the process of being commissioned, shut down, losing a further 737MW. Thirdly, Little Barford gas-fired power station, across the Bedfordshire border from Eaton Socon shut down in two stages, losing 885MW. These events all took place within a second of each other – 4:52:33-34pm.

Headlines such as ‘London Blackout Blamed on Drop in Wind and Natural-Gas Power’ probably gave the impression that the wind itself dropped and caused the loss of power, which was not the case.

Why did Hornsea and Little Barford shut down? We don’t know yet, but National Grid thinks they were both related to the lightning strike. Ørsted, the owner of Hornsea, has said:

‘During a rare and unusual set of circumstances affecting the grid, Hornsea One experienced a technical fault which meant the power station rapidly de-loaded – that is it stopped producing electricity’.

RWE, the owner of Little Barford, has said:

‘The gas station, which has a capacity of 730 megawatts, shut down in line with normal practice when demand hit 23,000 megawatts.’

Not sure I quite understand why it is normal to reduce supply when demand is high.

Why did so many people lose power? The electricity grid normally runs at a frequency of 50 hertz. The frequency falls if there is a drop in the supply of electricity, and if it falls below 48.8Hz, the system automatically curtails demand (ie cuts off people’s power, known euphemistically as ‘load shedding’). That is done by the next level down of grid operators, the distribution network operators (DNOs), rather than National Grid itself. All the DNOs cut some people off, as can be seen on page 21 of the ESO report.

One is left wondering why a lightning strike, a relatively common occurrence, caused such disruption, and why the loss of only two electricity generators had such far-reaching consequences. Also, if demand is to be curtailed, shouldn’t transport and hospitals not be the first to be switched off?

The government is looking into these issues, having instructed its ‘Electricity Emergencies Executive Committee’ (E3C) to produce an interim report by 17 September and a final one by 5 November.

My reading of the ESO interim report is that National Grid is only required by the regulator to have 1000MW of emergency power available to switch on in cases like these, and the losses here were more than that, which is one of the reasons that power cuts were necessary. One of the generation types that was brought online was pumped storage, where water is held back at a high level and released to produce hydroelectricity. The level of emergency power available is likely to be reviewed and increased – will that mean more generation being created just in case this happens again?

On page 25, the report reflects on the fact that the system has ‘lower inertia’ than it did before, ie new generation types such as wind and solar and more use of interconnectors from other countries can switch on and off quickly, whereas old generation types such as oil and gas take time to come onstream and go offstream, meaning there is less inherent stability in the system. ESO says it is considering ‘how new sources of inertia and other stabilising capabilities can be provided to the system ‘.

The report’s three ‘next steps’ are to understand better why the two generation sources switched off, review the ‘load shedding’ process with the DNOs and why some sources lost power in an uncontrolled fashion, and communications about such events.

‘Old Sparky’ in Private Eye has an interesting article (not online, sorry) that says National Grid’s job of balancing supply and demand is becoming impossible because of changes in generation types (the inertia issue above). ‘Microgeneration’, eg domestic solar panels, aren’t part of the system that National Grid can control thus aggravating the issue further. The article ends by saying that ‘the National Infrastructure Commission needs to develop its promised ‘analytical approach to resilience very quickly indeed.’

On the unexpected shutdowns, National Grid have concluded that the loss of power at Ipswich Hospital was a coincidence and was exacerbated by its reserve generator not initially switching on. Newcastle Airport was shut down because it was not registered as a ‘Protected Site’, but it has now been. Several trains that cut out couldn’t be switched back on again by their drivers and train manufacturer Siemens is looking at that issue.

I can come up with six lessons (to be) learned:

• find out why the lightning strikes’ disruption to the system caused an old-tech nearby power station to shut down as well as a new-tech far away one;
• check generators’ resilience to system shocks generally;
• review critical infrastructure and other sites that shouldn’t be switched off in situations like this;
• review how much and what backup electricity supply should be available;
• review whether the electricity system needs new methods to balance supply and demand given new generation sources; and
• review testing of backups and power restorations, not just for frequency but also effectiveness – apparently Ipswich hospital’s backup generator was regularly tested but still failed when a real power cut happened.

The ESO final report and E3C interim and final reports will be coming out in the next few weeks, let’s see what they have to say about these issues.