Six enormous concrete inlets, bored through the rock, drop 600m to turbines capable of generating 1,800MW of power. That’s around 5 per cent of Britain’s average daily level of demand.
Dinorwig Power Station was built in the 1970s to help back up the country’s electricity system. When power is cheap and plentiful, it pumps water from the valley to the upper reservoir, where it is stored as potential energy. The flow is released at times of sudden scarcity when prices are high.
Its concealment reflects the plant’s placement in Snowdonia National Park, one of Britain’s most picturesque mountain regions. “Our buildings above ground have to be faced with slate in the style of the old quarry buildings,” says John Armstrong, the station manager.
“The design had to be harmonious. It couldn’t detract from the local beauty of where we are.”
Dinorwig dates from when bigger generating units started to dominate the power system, as giant new nuclear and coal plants opened.
Its purpose was to protect the grid against demand surges or sudden outages. “Do you remember ‘The Thorn Birds’?” asks Armstrong, referring to the popular miniseries that aired on British television in the early 1980s.
Each episode caused a spike in electricity demand when everyone gathered in their homes to watch. “Some of the older guys remember running the whole plant flat out back then to meet the surge,” he says. “But that sort of thing doesn’t happen anymore.”
On-demand TV and more diffuse viewing habits have flattened the electricity curve, and television schedules rarely strain the system so severely anymore.
These days the calls are different. “Now when we open the sluices, it is generally because of the weather,” Armstrong says.Globally, storage used to be one of the electricity system’s backwaters.
No longer. As the world tries to cast off its dependence on fossil fuels, electricity grids are becoming ever more reliant on renewable generation, which needs backup for those times when the wind and sun don’t play ball.
Storage is one of the few green ways to do this, but the world has precious little of it. The technical and financial barriers to creating more are considerable, and policy makers are only beginning to grapple with these.
At last count, the world had the ability to store about 9,000GW-hours worth of electricity, almost all of it pumped-storage hydroelectricity like Dinorwig. Based on the 27,000TW-hours of electricity the world consumed in 2020, according to the BP Statistical Review of World Energy, that’s enough to cover just three hours of consumption.
Ubiquitous storage is key to expanding the reach of renewables and speeding the transition to a carbon-free power grid. As Bernadette Del Chiaro, executive director of the California Solar and Storage Association, put it, “Energy storage is actually the true bridge to a clean-energy future.”
Despite its big expansion of renewable energy in recent years, the UK’s storage needs remain relatively thinly served.
It has four pumped-storage hydroelectricity facilities (two in Wales and two in Scotland) that can store around 24 GW-hours of power and squirt it back into the grid at a rate of 2.3GW — meaning they could run at full power for around 10 hours before being exhausted.
Then there’s another 1.1GWh of battery storage, mainly linked to the country’s growing network of wind and solar farms. All told, that’s enough to keep the lights on for 45 minutes on an average day.
Britain will need more for its push to eliminate net carbon emissions by 2050, an objective accelerated by Boris Johnson’s government at November’s COP26 summit in Glasgow when he promised to decarbonize the country’s electricity grid by 2035 — a full decade and a half sooner.
Some question whether these goals are even realistic. According to Sir Dieter Helm, an energy economist at Oxford University: “There is no technology capable of backing up renewables at scale other than gas this side of 2035.”
Britain is presently transforming a grid dominated by “dispatchable” thermal power (such as fossil-fuel-fired turbines) into one driven increasingly by wind and solar generation.
While thermal stations can track consumer demand by flipping units off and on, renewables can only deliver what nature gives them. Solar, for instance, dies off in the evening just as demand peaks. This creates the need for backup, which is where storage comes in.
In 2009, David MacKay, a physicist then working for the UK’s Department of Energy and Climate Change, tried to quantify how much might be needed.
There are two key dimensions: how much power you can squirrel away and how quickly you can release it. “Think of it a bit like a watering can,” says Joe Worthington of battery maker Invinity Energy Systems Plc.
“The size of the can drives how much you can store, while the width of the spout determines how quickly you can put it back into the grid.”
MacKay looked first at short-term needs: the sort of intraday variability that might be caused by the sun going behind a cloud or the wind suddenly dropping, or weather changes from one day to the next.
He concluded that these swings were probably manageable; the main requirement was a large enough spout to meet demand hour by hour.
The much bigger issue was longer-term resilience: dealing with lulls when, say, the wind doesn’t blow for days on end. These are not uncommon.
In 2020, data from National Grid ESO, the UK’s national grid operator, and Exelon Corp showed that while wind generated up to 55 per cent of the UK’s electricity on some days, there were patches, including one five-day stretch and two of four days, where it produced 10 per cent or less. Such lulls require a very big watering can.
MacKay imagined a scenario in which the UK had 33GW of wind capacity (with an average output of 10GW), and a lull might last for five days.
The amount of capacity needed to replace all that lost wind output? A cool 1,200GWh — or nearly 50 times the capacity Britain has now.
MacKay died in 2016 before his study could be updated. Since then, the UK’s climate ambitions have widened. His hypothetical grid is far smaller than the one the country is now proposing.
To electrify transport and domestic heating, as its net-zero strategy requires, the UK will need much more generating capacity. The most heavily electrified scenario prepared by National Grid ESO envisages 157GW of wind on the network in 2050 (capable of delivering around 65GW on average).
Plug that number into MacKay’s calculations, and the storage needed just to cover that bit balloons to an astonishing 8,000GWh — almost as much storage as there is now in the world. And that’s not the full picture, either: It ignores another 90GW of solar and other renewables also expected to be on the grid.
In the nuclear and fossil-fuel electricity system that predominated when Dinorwig opened, large atomic and coal-fired power stations provided a constant “base load” of power to the electricity network. Peaks were met with open-cycle gas-turbine generators that could be run up to full power within minutes.
“People forget that we had a form of storage in those days,” says Michael Liebreich, chairman and CEO of Liebreich Associates, a clean energy consultancy, and senior contributor to BloombergNEF.
“It was called piles of coal sitting outside power stations. They were stores of energy waiting to be burned. By turning off coal, and now increasingly gas, we’re actually getting rid of about 70 per cent of the storage we had to back up the electricity system.”
The recent spike in UK energy costs has been exacerbated by a lack of fossil-fuel backup, following the decision to close the country’s last gas storage facility, at Rough in the North Sea, in 2017.
It’s only because so much fossil-fuel infrastructure still exists that Britain hasn’t yet had to build more electricity storage. That is despite running a grid with 36 GW of wind and solar capacity on it — larger than the hypothetical one that MacKay believed would need thousands of gigawatt-hours of storage to back up.
But with Britain’s fossil-fuel-generating capacity set to fall by a quarter this decade, before vanishing by 2050, the question is how much of the growing gap can be filled with batteries, pumped storage and the like.
At present, almost all electricity storage is pumped hydro, accounting for around 98 per cent of global capacity, according to the International Renewable Energy Agency. The rest is mainly lithium-ion batteries — the sort used in electric cars.
Pumped hydro plants such as Dinorwig can deliver lots of power in a matter of seconds, and storage depends only on the capacity of the reservoir. However, the 2.3GW total output of the UK’s four existing plants is nowhere near a big enough spout to cover the 65 GW of average wind output alone on the grid in 2050 that could go missing.
Nor is the total energy stored anywhere near enough.
Try to build more, though, and you quickly run into constraints. Only a certain number of sites in the UK have the right characteristics, mainly places of natural beauty.
The utility SSE Plc is planning a 30GWh project at Coire Glas in the Scottish Highlands that would more than double the UK’s capacity, as well as add 1.5GW to output.
But further developments are likely to attract fierce environmental opposition. Dinorwig had to be put inside a mountain to get around such objections. MacKay estimated that the maximum might be no more than 400 GWh.
Battery use has been soaring as its cost per kilowatt-hour of storage has tumbled, thanks mainly to innovation led by the makers of electric cars. According to BloombergNEF, the price per kWh of lithium batteries has fallen by 89 per cent in real terms since 2010 to US$132 per kWh.
Batteries work well for storing relatively small amounts of energy.
The problem comes when you try to do more. “Batteries do not scale as well as other technologies because the energy conversion and storage systems are linked,” says David Cebon, professor of mechanical engineering at Cambridge University.
“There is no significant reduction in cost per KWh with increasing storage capacity. Each additional kWh costs about the same as the first.”
To get a sense of the difficulty, try covering that 8,000GWh shortfall with lithium-ion batteries and — even at US$132 per kWh — the total cost would be an astonishing £741billion (US$1trillion), nearly 40 per cent of UK gross domestic product.
Then there are other issues, such as batteries’ consumption of scarce rare metals, as well as the steady degradation of batteries after a certain number of charge and release cycles.
Put it all together, and few believe that batteries will have a substantial role in providing backup for longer lulls or cover for cold winters.
BloombergNEF estimates that, for all their growing adoption, the UK’s battery output will total just 9.7GW by 2030, with a storage capacity of less than 30GWh.
How to make up the rest is a question that is energizing entrepreneurs.
Among their more heavily touted options are those that propose using air as a storage medium, either compressed and forced into disused salt caverns, or liquified and held in metal storage tanks at very low temperatures.
When the power is needed, the air is released through turbines that drive electricity generators.
Compressed-air technology has existed for decades, but only two plants are in operation — one in Germany dating from 1978 and another in the US, which opened in 1991.
Adoption has been feeble mainly because of its lack of “round trip” efficiency — a key storage metric that measures how much of the energy used in the process is recovered at the other end. As storage is essentially funded by arbitrage — buying cheap power at times of abundance and selling it at a premium when electricity is scarce — the more you lose in the process, the higher that sale price has to be.
Compressed air uses lots of energy to squash the air on the way in, while reversing the process requires the (now very cold) expanding air to be reheated — generally with gas — before going through the turbines. Not only is more than half the energy lost, but the gas reheating creates emissions, too.
Perhaps closer to commercial acceptance is liquid air storage, which liquifies air by freezing it and then stores it in steel tanks, thus both obviating the need for geological storage and making it more easily scalable. “Need more storage?” says Cebon.
“Simply add a few more tanks.”
Liquid air’s efficiency is around 60 per cent, making it quite expensive, although this can be increased by introducing heat into the process.
London-based Highview Power Storage has built a 15MWh demonstration plant near Manchester.
Having raised some US$145 million in funding and grants, it recently started building its first commercial facility — a 250 MWh facility that can nominally supply 50MW for five hours.
“Liquid air looks like the most promising technology,” says Cebon. “Unlike others, it has a high readiness level and doesn’t suffer from any natural or geographical constraints.”
Move beyond these and you drift into more eccentric-sounding territory — schemes to flood old coal mines with hot water, build giant underground flywheels or machinery that lifts concrete blocks and then drops them, using gravity to make electricity.
Another proposes running railway cars loaded with concrete up hills, but this again is nature-dependent and somewhat ill-suited for British terrain.
The biggest concern is the lack of a final line of backup against a prolonged winter wind drought. Such weather patterns can afflict wide geographical areas, occurring once every two to five years.
Northwestern Europe experienced a nine-day dunkelflaute, a German phrase for a cold snap that creates high demand for electricity with little to no power able to be generated by sun or wind, in 2017.
Few affordable options exist to meet such contingencies. One idea is to make hydrogen by diverting surplus renewable power from windy or sunny days through electrolyzers to produce a “green” version of the gas.
This could then be stored in caverns and burned to produce electricity when other stores are exhausted.
The main problems? Well, green hydrogen is very expensive, costing from US$2.50 to US$6 a kilogram to make (equivalent to US$112.50 to US$270 per barrel of oil).
And that would rise further if the electrolyzers were only used part-time. Then there is the round-trip inefficiency of the process — just 30 per cent of the energy survives the journey — which would push up the price further.
And lastly, there is the sheer amount of renewable capacity that would be necessary. “You would also need to fill the North Sea with wind turbines to make enough green hydrogen to make it work,” says Cebon.
Britain is unlikely to ever have sufficient pure storage to meet every eventuality. To cover a theoretical three-week dunkelflaute would take 33,000 GWh of storage just to make up the lost wind part, based on MacKay’s calculations.
National Grid ESO’s plans for 2050 take into account just a fraction of those numbers. Instead, it has turned the problem on its head, looking to fit demand around the uncertainties of production.
“You can get away with much less storage if you can manage demand as well as supply,” says Julian Leslie, head of networks at the grid operator.
National Grid ESO expects peak demand in 2050 to hit around 140GW.
But “demand response” measures could reduce that to just 79GW — a decline of 43 per cent — through interventions. These include paying industrial and commercial customers to switch off at times of stress, and also getting consumers to cut or balance out their usage — for instance, charging cars or storing heat during off-peak hours.
There are also hopes that electric vehicles themselves could provide storage. Assume an average battery size of 40 kWh and 30 million vehicles on the road, and you can theoretically store 1,200GWh (assuming, of course, that you can reliably draw it down in times of need).
Yet all these ideas assume that consumers will respond reliably to incentives and that technology will exist that can draw electricity back from their cars, or seamlessly curb usage without changing their experience.
The grid is hedging its bets. It also expects to rely on fossil-fuel generation — abated with carbon capture and storage, a technology that has yet to be proven at scale and for which Britain has no policy framework in place. Hopes are also being pinned on securing large-scale imports through interconnectors to other countries.
While these hedges could be scaled up, they introduce other uncertainties. In the case of interconnectors, there’s the risk of failure. Britain gets just 5 per cent of its electricity through these links, but when a fire recently severed one of the cables across the English Channel to France, it prompted fears of winter shortages.
Meanwhile, French President Emmanuel Macron threatened to cut off supplies to the island of Jersey over a dispute about fish — an eloquent demonstration of political risk. Connectors also depend on other countries having surpluses to share.
“What happens when a whole bunch of interconnected countries all experience shortages at the same time?” asks Peter Atherton, an independent energy analyst. “Will these fancy links really work when we need them?”
Storage isn’t cheap.
When Dinorwig was built 40 years ago, it cost around £400 million, or around £2.5 billion (US$3.39 billion) in current terms. That’s equivalent to capital expenditure of around £1,400 per KW of output.
The cost of electricity from wind turbines and solar panels may have fallen dramatically, but storage represents an “on-top” charge that consumers would have to assume.
A government report in 2020 warned that as renewable penetration increases, the cost of its intermittency could increase dramatically.
There is already a “capacity market” that offers financial subsidies to companies supplying short-term storage to stabilize the grid. Payments are fixed by auction and linked to the amount of storage offered. The UK government recently launched a consultation process looking at longer-term storage.
Many think capacity payments will be needed there, too.
Observers worry that this ad hoc splashing of subsidies may lead to an encrustation of unnecessary costs for consumers.
“It would be much better if we had a robust market mechanism for the whole energy market rather than one special CFD (contract for difference) for this technology and another one for that,” says Guy Newey, director of strategy and performance at the Energy Systems Catapult, the non-profit set up by the government in 2015 to speed the transformation of the UK’s energy grid.
Tim Stone, chairman of the Nuclear Industry Association, says grid operator ESO, or some other competent body, should be given a formal duty to consider the all-in costs of delivering a green-energy network.
“We need a system architect sitting above it who is thinking how we can do this for the lowest cost to the national economy,” he says.
No one doubts the need for storage in a zero-carbon grid. But the amount potentially required is massive. How much is viable is ultimately an economic or even a geographical question.
Globally, storage is concentrated in a relatively small number of developed markets, such as the US, the UK, France, Australia and Japan.
China and India are also making big investments. Yet aside from a few — generally sparsely populated — countries blessed with abundant natural resources, such as Iceland and Norway, most countries are far from having the resources to back up anything close to a fully renewable grid.
The US, for instance, has the ability to store 595 GWh, according to the Department of Energy. That’s roughly an hour of US demand.
Given its natural constraints, the UK faces some hard choices.
One option might be to build a less heavily renewable grid, with a higher share of nuclear power — the only dispatchable zero-carbon source the world now possesses. In some instances, Britain might even fall back on fossil fuels to meet any shortfall.
According to Cebon, that might be the only rational way to deal with those winter dunkelflautes, assuming a renewable-dominated grid.
“For the occasional two weeks when there is no wind in the middle of winter, do you really want to build a whole new hydrogen infrastructure to deal with that?” Better to top up the system with natural gas, for which the plants already exist, thus obviating the need to build new infrastructure.
“Burning a bit of gas every few years isn’t such a big deal,” he argues. “If that’s all you are doing, I would say it’s essentially problem solved.”