2 days ago
Britain's electricity grid is dangerously outdated
Built for a different era, Britain's electricity grid is now creaking under the weight of the energy transition, and no one is fully in charge of fixing it.
Some of the equipment still in use dates from the 1950s. And the software that runs the grid? That was written before the current electricity market even existed.
Both physical and digital infrastructure are suffering from decades of underinvestment as the connection of renewables has been prioritised above all else.
Evidence shows substantial legacy equipment installed between 1955 and 1975, much of which remains in service.
Last year, a report for energy regulator Ofgem by Cambridge Economic Policy Associates (CEPA) found that a third of transformers and 30pc of switch gear were built in the 1970s, as were just over half of all power cables.
The transformer that blew up at North Hyde in March, causing the Heathrow blackout, dated back to the 1960s.
More than a third of other grid equipment, from protection relays to communications systems to fire suppression, was built before the 1970s.
The same CEPA report indicated that some of the grid's most important control systems typically only last 10 to 20 years before needing replacement, and that traditional network assets (overhead lines, underground cables, transformers, and switchgear) generally last between 41 and 70 years. This suggests a lot of grid assets are pushing the boundaries of their expected lives.
The obsolescence risks associated with modern smart grid infrastructure are higher. The report notes these devices 'are often tightly integrated with software and specialised equipment that quickly become obsolete ... The typical software and equipment lifecycle of 10 to 15 years can limit the technical lives of these assets'.
The fragility of the digital infrastructure was highlighted recently. May 29 was a day of high wind output, but there was a lot less wind than expected the day before, which is when most power trading takes place. Lots of exports on UK interconnector cables were booked, and not all could be reversed on the day.
The problem was exacerbated by the fact that until 10am, the within-day wind forecast was significantly higher than the already inaccurate day-ahead forecast. At its height, the difference between actual wind output and the forecasts was 4.5 GW.
This represented 28pc of wind output at the time and 17pc of demand, a huge discrepancy for the National Energy System Operator (Neso) to manage.
That day, the Neso's control room had to implement almost 25,000 balancing actions, turning down wind and tweaking gas power stations' output in an effort to keep the grid running properly.
All this was done on software built in the 1980s. A literal relic of the days of privatisation, this software pre-dates the current market structure, known as NETA (New Electricity Trading Arrangements), which replaced the original post-privatisation structure in 2001.
There are increasingly frequent outages on the system that connects the Neso control room with the control systems of power stations. When this happens, Neso staff must revert to notifying their instructions by telephone. Some 25,000 balancing actions in one day represents 17 a minute – clearly not something the small control room team can do manually.
Had the software gone down that day, keeping the lights on would have been a matter of prayers as control room staff would have had to revert to a simpler way of doing things: getting wind off the system, bringing up all available gas plants, and, if necessary, suspending exports.
Doing this would have made balancing by phone feasible, and would have removed a major source of variability that challenges system balance: wind. Not only can wind output swing dramatically, Neso uses inadequate forecasting tools that are widely criticised within the industry for their poor accuracy.
A major project to replace the ancient balancing system, begun in 2014, was cancelled last year after a decade of failure, with no visibility on when the new replacement will be operational.
Neso recently launched an audit of its demand forecasting, indicating its models had not been reviewed in '10-20 years'.
That is despite the significant changes represented by the introduction of renewables, many of which are connected to low-voltage grids over which Neso has little visibility (and which appear as negative demand in Neso's systems).
Many parts of the grid still rely on control systems developed decades ago, well before the smart grid era.
One such system is SCADA (Supervisory Control and Data Acquisition), which allows operators to monitor and control grid equipment remotely. In some cases, these systems are running on obsolete software like Windows 3.1 or early versions of Windows XP, long out of support and riddled with known vulnerabilities.
In power stations, these controls are sometimes used for synchronising generators with the grid, a safety-critical function. These outdated systems are far more susceptible to cyberattacks, hardware failure and compatibility issues with modern infrastructure.
The Grid Code – the rulebook for the electricity system – says remarkably little about digital control systems. There's no mention of software integrity, cybersecurity or the use of obsolete platforms like Windows 3.1.
As demand on the grid increases and the system becomes more complex, continuing to rely on decades-old physical and digital infrastructure represents a growing threat to resilience.
Yet, shockingly, no one has responsibility for ensuring industry rules will guarantee a secure system, leaving us in the hands of the original architects of a grid that bears little resemblance to today's power market.
The question is no longer if the system is fragile, but how long it can survive without a major failure.
Broaden your horizons with award-winning British journalism. Try The Telegraph free for 1 month with unlimited access to our award-winning website, exclusive app, money-saving offers and more.
