We use cookies to to enhance the service we deliver you. By using this site, you agree to our use of cookies as described in our Cookie Policy.

Skip to main content
News Regulation Supply Networks

Blackout investigation: What went wrong at Hornsea One and Little Barford

Image: Getty.

Image: Getty.

Ofgem's investigation into the 9 August blackout has detailed why Hornsea One and Little Barford - the offshore wind farm and CCGT plant at the heart of the event - remained disconnected from the grid.

As a result, Orsted and RWE have agreed to make a voluntary payment of £4.5 million to Ofgem’s redress fund.

The investigation has also made recommendations for the distribution network operators (DNOs), as well as shed light on the role of distributed generators in the blackout and the need for greater visibility of the assets.

Hornsea One

At 16:52:33, a lightning strike caused a fault on the Eaton Socon – Wymondley 400kV line. Whilst this was rectified within 80 milliseconds, around 150MW of distributed generation disconnected from the local distribution networks due to a safety mechanism known as vector shift protection.

At Hornsea One, the onshore control system operated as expected during the incident. However, the offshore wind turbine controllers reacted incorrectly to voltage fluctuations on the offshore network. This caused instability between the onshore control system and the individual wind turbines, and its this instability that triggered two modules to shut down, with Hornsea One deloading from 799MW to 62MW.

In Orsted’s internal investigation, it identified that the stability issue had occurred around ten minutes prior to the incident on 9 August but had not caused deloading at that time.

Orsted’s also provided Ofgem with information on its modelling prior to 9 August, which showed problems with the voltage control system when operating at its full capacity of 1,200MW. These findings were not shared with the ESO at the time. Plans were in place for a software update scheduled to take place on 13 August to resolve this issue, with the update then being implemented on 10 August following the blackout.

Orsted also did not notify the ESO when it deloaded by 737MW, and temporarily began its process of starting up two of its modules without coordinating with the ESO.

Hornsea One has acknowledged that it did not meet its Grid Code requirement to remain connected and transiently stable following a fault on the transmission system, with power output recovering to at least 90% within 0.5 seconds, having deloaded following the fault.

In addition, it has accepted that it did not meet the Grid Code requirement to have an overall voltage control system that appropriately dampens or limits swings.

Little Barford

Within a second of the fault, Little Barford’s steam turbine, which was generating 244MW, tripped. The trip occurred initially because of discrepancy in the three independent speed sensors on the turbine, which exceeded the tolerance of the control system. However, the root cause of the discrepancy has not been established.

This resulted in a total loss of generation of over 1,1039MW within one second of the fault. This caused the frequency of the electricity system to fall at a rate of change of frequency (RoCoF) above 0.125Hz/s, which then resulted in an estimated 350-430MW of distributed generation tripping off unnecessarily, according to the investigation.

Frequency response was then activated, with frequency fall stopped 25 seconds after the fault at 49.1Hz, plateauing after 45 seconds at 49.2Hz, below the minimum frequency of 49.5Hz set in the SQSS.

However, Ofgem criticised the frequency response as "inadequate", with primary responders under-delivering by 17% and secondary by 14%. Despite this, Ofgem doesn't believe better response and reserve delivery would have stopped demand being disconnected.

Mandatory response providers and commercial Fast Frequency Response providers of dynamic primary response under-delivered by approximately 25%.

Around a minute after the fault, a gas turbine generating 210MW at Little Barford was shut down for safety due to too much steam pressure in its pipework.

The second gas turbine at Little Barford generating 187MW was manually tripped by plant staff around a minute and a half after the initial fault due to safety concerns.

RWE Generation - which owns Little Barford - has acknowledged the role it played in contributing to the power outage by not continuing to provide power to the system following the fault, Ofgem said in its report.

Distributed generation and the DNOs

Ofgem’s investigation also found an estimated distribution generation loss across the event between 1,300MW and 1,500MW. At least 500MW was lost due to the loss of mains protection settings in the first second of the event, and over 200MW tripped when the system reached 49Hz.

When the system frequency dropped below 48.8Hz, the DNOs disconnected 892MW of net demand in a process known as Stage 1 of Low Frequency Demand Disconnection (LFDD).

The ESO reported that the net demand reduction seen by the transmission system was 350MW, meaning around 550MW of additional distributed generation was lost.

While most DNOs met requirements regarding LFDD, there were some issues with essential services such as trains and hospitals being disconnected.

However, it is difficult to isolate sites and so those sites should have their backup generation in place.

There was also some “concerning” evidence that some DNOs disconnected generation via LFDD that was providing frequency response or reserve services.

Ofgem also concluded that DNOs don’t collect and record enough data on distributed generation, which highlights the “substantial improvements” needed for DNOs to transition into DSOs.

More granular data collection is needed, as the majority of data the DNOs provided Ofgem only offered a partial view of which distributed generators tripped.

The regulator is continuing to review the behaviour of distributed generators during the event, and will consider the appropriateness of opening investigations into any licensed parties’ compliance with Distribution Code requirements regarding distributed generators’ protection settings.

Recommendations for the DNOs and distributed generation:

  • The ESO and DNOs should review the timescales for the Accelerated Loss of Mains Change Programme and consider widening its scope to include other distributed generation that unexpectedly disconnected or deloaded on 9 August. This should be done through the Energy Networks Association (ENA), which should put forward its recommendations to the E3C by April.
  • Ofgem and BEIS should undertake a joint review of the regulatory compliance and enforcement framework for distributed generators, engaging with the industry in Spring 2020.
  • The E3C, through the DNOs and ENA, should undertake a fundamental review of the LFDD scheme, reporting its progress to Ofgem and BEIS on a quarterly basis.
  • Ofgem should consider options to improve the real-time visibility of distributed generation to the DNOs and the ESO.

Loading...

End of content

No more pages to load