Automatically Averting Blackouts with GridDuck

On the 9th of August 2019 a large power shutdown occurred, leading to the disconnection of over a million customers and various infrastructural systems. Commuters were left stranded during the Friday evening rush hour, with some forced to use mobile phone torches to exit the London underground. Additionally, the failure of a backup generator left Ipswich hospital without working lifts, X-rays, and other non-battery-secured systems.

The cause of this was determined to be the near-simultaneous failures of two power plants, both coincident with a lightning strike on a power line in Hertfordshire. The power plants affected were a gas generating plant in Bedfordshire, and the Hornsea offshore wind farm. Their failure meant that there was a fall in the frequency of the grid, triggering the launch of back-up systems and the disconnection of power to certain areas.

National Grid’s systems responded immediately and apparently as planned. While they were successful in stopping the frequency fall caused by the shutdown of the first plant, the loss of the second meant that the deficit exceeded the 1GW of redundancy required by the Security and Quality of Supply Standards (SQSS). As a result, systems in the low frequency demand disconnection (LFDD) scheme were cut off from supply.

Future Risks

Such shutdowns occur occasionally in isolation, where there is usually sufficient backup capacity to compensate. The last cut of this scale was in 2008, which occurred due to the failure of a coal-fired power plant and a nuclear plant. We can minimise shutdown risk by increasing backup capacity (and therefore energy prices) or focus on building smaller generators, which cause less stress on the grid and make coincident failure much less likely.

It should be noted that these actions will never eliminate the possibility of a shutdown. Additionally, the reliability of renewable resources is always going to hinge on weather conditions. This reliability has traditionally been ensured by aspiring to have an excessive number of sites present in different areas. The wind doesn’t always blow everywhere, but usually blows somewhere. However, so far there is not nearly enough redundancy in this system to prevent shutdowns and, for a company that aspires to become carbon neutral by 2050, there has so far been little provisioning to secure the reliability of the grid.

We therefore have three options to prevent shutdowns: first, increase back-up capacity in the form of batteries, hydroelectric, or nuclear power plants; second, build a system of renewable sources that far exceeds the demand of the grid; or third, find a way of containing the impact of power outages by more selectively allocating disconnections.

Targeted Disconnections

The Ofgem report on the shutdown advises that the list of facilities connected to the LFDD scheme should be reviewed to ensure no critical infrastructure is disrupted. Alternatively, national grid could endeavour to directly shut off nonessential appliances, thereby not imposing blackouts on consumers or infrastructure.

Until recently, a high cost for such a scheme has left it off the cards. However, recent experience at GridDuck has shown that this flexibility could be built in for a start-up cost of around £500 million, then £20–30m per year. It should be noted that Ofgem has the power to fine 10% of National Grid’s turnover for the recent shutdown, equal to £1.525 billion.

Three interruptable appliances that have been the subject of GridDuck testing have been identified as good candidates for a shutdown system. Together, refrigeration, air conditioning, and heating systems in businesses account for 10% of total UK demand, so shutting off half of them would have been sufficient to make up for the 1.7GW deficit of the recent power failure. A fail-safe system of this kind would be reliable, would not contribute to energy price increases, and would not leave consumers in the dark.

The Smart Grid

In an industry required to shift to renewable and more variable sources of energy, there are significant challenges which are so far not being addressed. The lack of provisioning to secure set, reliable capacity buffers has led many commentators to suggest that demand side response would be the most efficient and elegant solution to the shutdown problem. Whatever the case, action is required to compensate for the near half of nuclear energy capacity due to be retired by 2025, and the simplest course may be to tie emergency shutdown systems to the rapidly emerging smart grid.

If you’re interested in getting involved with the Smart Grid check us out at https://gridduck.com