How renewables caused Scottish grid’s ‘double heart attack’

In December 2017, a National Grid report ‘Performance of Phase-Locked Loop Based Converters’ [PDF] highlighted concerns about the ability of the billions of pounds of investment in HVDC interlinks being installed across the UK to transfer power in all circumstances. Since then, I have been questioning whether power companies – including those in Scotland – have solved the known instability risks.

National Grid identified that these could emerge as a result of progressively increasing the use of renewables while simultaneously closing synchronous generation. Hunterston B nuclear power station’s last reactor, No 4, ran for the final time in January this year, further weakening the Scottish grid.

The National Grid report concludes that “system strength will decrease in our transmission network over the next decade” and that “We have found an increasing risk of converter instability”. It goes on to predict that as system fault levels fall due to increased reliance on wind turbine generation, when network faults occur there will be many scenarios where the DC/AC converters will become unstable and voltage will start to surge and oscillate.

In a presentation hosted by the Global Power System Transformation Consortium and Energy Systems Integration Group at the end of November last year, National Grid Electricity System Operator head of networks and chief engineer Julian Leslie reported that Scotland’s grid suffered two ‘heart attacks’ (my metaphor) during the night of 24 August 2021.

If there has been a public statement on this crisis event from Scottish and Southern Energy, Scottish Power or EDF, the owners of Hunterston and Torness, I have missed it. Graphical records presented show “severe voltage disturbances” lasting between 20 and 25 seconds (a long time in the context of power system engineering) on two occasions, 30 minutes apart. Oscillations for the 400kV system ranged between approximately 355kV and 435kV with a frequency of around 8Hz, with the most significant close to the northern end of the £1.1bn HVDC link at Spittal, south of Thurso. Huge power oscillations on a similar scale would have been seen at various locations.

The presentation identified that “Some users tripped off during the disturbances”, which, I understand from other sources, apparently included several wind farms. We appear to have come close to losing most of the grid north of the Scottish central belt. Coincidentally, the ageing gas station at Peterhead that would have provided much stabilising synchronous generation was reportedly offline at the time.

From discussions within the profession, a suspected potential trigger may have been the newly commissioned Moray East offshore wind farm interacting with the HVDC Spittal DC/AC converter nearly 50 miles away within the very weak north Scotland grid. In addition, the disturbances apparently reached Torness nuclear station near Dunbar, just east of Edinburgh, considerably further away.

It is unclear what saved us, and it is a concern for the profession that Torness was adversely affected and likely came close to tripping. This would have had very serious consequences for the central belt, home to half of Scotland’s population.

Detailed investigative work being carried out by National Grid ESO includes an extensive joint examination of the Scottish Power and SSE network capability. A working group will investigate areas such as electromagnetic transient modelling of the supergrid in the north of Scotland as converter-based generation increases, and will secure all relevant data on network configuration and numbers of turbines in operation etc before recommending remedial actions.

It is possible Ofgem might also decide to take a close look at the incident, especially as progressive weakening of the grid in Scotland through its increasing of renewables will likely be creating an increasing risk of periodic unpredictable effects on the rest of the UK grid.

Now Hunterston has closed, it leaves a significant drop in system strength (short-circuit level) in West Scotland. A consequence is that the Wales/Hunterston HVDC link, which is not of VSC design so not capable of supporting black starting, could in future encounter fault conditions that lead to instability. In that situation, Scotland could not use it to import and would have to disconnect.

With Torness scheduled for closure in 2028, in the period between November 2021 and its shutdown Scotland will have lost 2,000MW of stabilising nuclear generation. Without urgent action, a further weakened Scottish network will possess only the ageing gas-fired Peterhead power station, some small-scale pumped storage and hydro plus, in vast predominance, intermittent renewables. To ensure best available dependability when renewables routinely collapse, it will have to rely on existing overhead line imports from England ahead of HVDC.

While National Grid ESO in its 2020 ‘Future Scenarios’ document [PDF] identifies “very strong” and “strong” strategic requirement “need cases” to provide short-circuit reinforcement at eight Scottish sites including Hunterston, plus increased inertia across Scotland to protect frequency, all such provision comes at huge capital cost, and is a direct consequence of us increasing renewables. The present scheduling for its provision appears to run from autumn this year until 2034 which, crucially, seems to be at serious odds with the required urgency.

My question about stability was demonstrably answered in August last year, however. Until the event is fully understood it can happen again tomorrow and could be catastrophic.

David B Watson is a chartered electrical engineer who before retirement was manager of projects at Foster Wheeler Energy, based in Glasgow and responsible for project-execution management at the company’s Scottish operation.