At 5pm on Friday 9th August, nearly 1 million people were affected by fairly widespread power outages across England and Wales. National Grid have stated their systems were not to blame but some believe high levels of wind generation could be to blame. The Energy and Emergencies Executive Committee have been tasked to investigate whether National Grid’s procedures are fit for purpose by Business Secretary Andrea Leadsom as Ofgem request an urgent detailed report to find out what went wrong.
National Grid has an obligation to maintain the UK’s frequency between 49.5 Hz and 50.5 Hz. However, on Friday the frequency dropped sharply to a low of 48.91 Hz as a gas-fired power station at Little Barford, Bedfordshire failed at 4:58pm and a mere two minutes later, Hornsea offshore wind farm also disconnected. National Grid was unable to call enough generation on and Frequency Response services couldn’t avoid the frequency dropping to the above levels where demand had to be lowered automatically by initiating power cuts to demand centres deemed non-essential. National Grid’s Director of Operations Duncan Burt explained that the near-simultaneous loss of two generators was more than the grid was routinely prepared for. The last comparable event in the UK happened over 10 years ago as Sizewell B units tripped and almost caused widespread power outages as several coal fired power stations almost tripped as a result, showing this is a relatively rare and unfortunate incident for the grid.
On Friday, demand was lower than recent days and the tail off for the day happened slightly earlier too which is normal for the end of the working week. As demand lowered, National Grid demand forecast error for settlement period 35 increased, most likely due to disconnections. The forecast demand errors turned negative, meaning outturn demand is below forecasted demand. This combined with high, but by no means record, wind levels and slightly higher nuclear production (which should be increasing in the coming days) meant Combined Cycle Gas Turbines (CCGT) were pushed from the generation stack to run below recent levels. At this point, 3 units tripped, firstly Little Barford CCGT (LBAR-1 in Bedfordshire) then 2 of 4 Hornsea offshore wind farm units (HOWAO-2/3) off the coast of Hull. It’s unlikely the Hornsea issue was linked to high wind speeds as this would usually cause clusters of turbines to trip rather than entire arrays.
It can be argued that the impact of the outage was limited on price as the event was a system problem. It’s clear that the trips pushed intraday prices up with a 10 GBP move but the price did stay within the range of the week. The market was already firming up at this point and whilst the trips undoubtedly gave it a helping hand, the price movement wasn’t as drastic as some would expect (see Figure 1). With this said, the whole day still delivered well above day ahead (DA) Auction levels at cash-out. To put this price movement into context, at 21:00 two days later system prices were almost £15/MWh higher than during the power cuts. All in all, clients on cash-out products would have experienced higher prices during this time but there will be no impact for clients on fixed energy products and clients looking to secure a contract in the coming days most likely won’t find the wholesale market has shown much volatility directly related to this event.
As the ancillary markets have become more competitive with demand side response (DSR) services and open-cycle gas turbines (OCGT) muscling in, the traditional plants to react to this type of event, Dinorwig Power Station (DINO), Ffestiniog Power Station (FFES) and Cruachan Dam (CRUA) pumped storage, are largely unavailable to National Grid in useful timescales. Instead, OCGT’s and aggregator units likely get called upon and these units will be doing little to solve the inertia problem National Grid have. Maintaining system inertia is expected to pose another serious challenge for the evolving, increasingly decentralised UK power system. It will be interesting to see whether this prompts any change in policy from National Grid, potentially more spinning reserve on days like this which will likely increase price volatility and BSUoS rates. It is also worth mentioning that National Grid are well aware of these potential problems and have published documentation on their commitment to running a safe, stable and zero-carbon grid by 2025. They plan to do this through market mechanisms for frequency management, inertia and short circuit infeed, and voltage management as well as increased situational awareness.
In conclusion, a very difficult day in the office for National Grid. They were already offering up quite a bit of CCGT generation but possibly in hindsight, some wind generation needed to be taken down (at hugely negative prices) and even more CCGT’s offered up. It is tough to be too critical in this situation given the wind levels weren’t the highest we’ve seen, the demand wasn’t that low, and the trips weren’t something you would expect to see several times a week, let alone within 2 minutes of each other.