Halfway up a slate mountain in Snowdonia, hidden inside a cavern the length of a football pitch, six turbines sit in the dark waiting for a phone call from the National Grid. When the call comes, water begins falling through a shaft carved into the rock, and within seconds the machines spin up to deliver more than a gigawatt of electricity to homes across Britain. The station is Dinorwig, and it is one of the fastest responding power plants ever built.
The scale of what happens inside the mountain is difficult to picture from the outside. Elidir Fawr, the peak that houses the plant, looks like any other Welsh summit ribbed with old quarry scars. But its interior has been hollowed out. A reservoir sits near the top. A second reservoir sits at the bottom. Between them: a substantial vertical drop and enough plumbing to move a small lake.
How a mountain becomes a battery
The idea behind pumped hydro is old and simple. Push water uphill when electricity is cheap and abundant. Let it fall back down when electricity is scarce and expensive. The reservoir is the battery. Gravity is the discharge circuit.
At Dinorwig, the upper reservoir holds a volume of water reported by the operator to be in the millions of cubic metres — enough that a full drawdown can sustain generation for several hours before the top pond runs dry. Overnight, when demand collapses and baseload plants have surplus power, the turbines reverse and run as pumps, hauling the water back up the shaft to be used again the next morning.
The economics only work because of the price gap between night and day. Cheap electrons in. Expensive electrons out. The round trip loses energy to friction, heat and turbulence, but the arbitrage still pays.
From standstill to a gigawatt in seconds
What makes Dinorwig unusual is not the storage. It is the speed. The plant was designed in the 1970s specifically to catch the sudden surges in demand that hit the British grid at predictable moments — the end of a popular television programme, the whistle at the end of a football match, the ad break during a soap opera. Millions of kettles switch on within seconds of each other. The grid needs power immediately, not in ten minutes.
Dinorwig was built for that job. The turbines can be brought from a spinning standby condition to full output in a matter of seconds, and the plant can reach full capacity remarkably quickly. That is faster than any gas plant, faster than any coal plant, faster than almost anything else on the system apart from an interconnector already carrying current.
The comparison worth holding in your head: a large modern combined-cycle gas turbine takes half an hour or more to reach full output from cold. Dinorwig does it in the time it takes to boil the kettle it is powering.
Why Britain built it in the first place
The station was commissioned in the 1980s at significant cost, and it was one of the largest civil engineering projects in Europe at the time. Many kilometres of tunnels were bored through the mountain. Large volumes of rock were removed. The main machine hall, carved out of solid slate, is tall enough to hold the nave of a cathedral.
The reason for all that engineering was simple: the grid of the 1970s was dominated by coal and, increasingly, nuclear. Both are slow to ramp. Both prefer to run flat out around the clock. Neither can respond to a five-million-kettle spike. A fast-acting reserve was needed, and hydro was the only mature technology that could provide it at gigawatt scale.
Four decades on, the same problem exists in reverse. Coal has largely gone. Wind and solar dominate new build. But wind and solar cannot be dispatched on command — they arrive when the weather says so — and the grid still needs something that can absorb a surplus at three in the morning and hand it back at six in the evening.
The renaissance of pumped storage
After decades of neglect, pumped hydro is having a moment. The British government has ended a long freeze on new large-scale storage. In Gwynedd, the French utility ENGIE has announced the next stage of a £1 billion investment programme in its Welsh hydro stations, including Dinorwig and its smaller sister plant Ffestiniog.
The renewed interest is driven by mathematics. Batteries are excellent for shifting energy by a few hours. They struggle to shift energy by a few days. Pumped hydro scales in a way lithium cannot. Recent analysis of global pumped hydro potential suggests there are hundreds of thousands of suitable off-river sites worldwide, with combined storage capacity sufficient to support large-scale renewable energy deployment.
The point is blunt. Pumped hydro is more than a century old, uses no rare metals, and can last far longer than chemical batteries.
Australia is building the next Dinorwigs
The clearest sign of the revival is on the other side of the world. Australia is pouring concrete on a generation of pumped hydro projects that dwarf what Britain built in the 1980s. In New South Wales, planning minister Paul Scully has approved a pumped hydro and solar project at Gloucester, on the site of a former coal mine. The plant will pair a pumped hydro station with a solar farm, using the panels by day to lift water into an upper reservoir carved out of the bushland.
It is the first pumped hydro project approved in the state since Snowy 2.0, the vast expansion of the existing Snowy Mountains Scheme. Snowy 2.0 offers substantial storage capacity over a design life of many decades. A large utility battery, by comparison, delivers a few gigawatt-hours at best and expects to be replaced within two decades.
The scale is what matters. Large-scale storage is essential to supporting high levels of solar and wind penetration. Britain today has relatively limited pumped hydro capacity, most of it in the two Welsh stations and a handful of Scottish sites. The gap is what the Gwynedd investment is meant to start closing.
The physics of a multi-hour battery
The energy stored in a reservoir is embarrassingly simple to calculate. Mass times gravity times height. Every cubic metre of water lifted through a substantial height stores potential energy, before losses. Multiply by the volume of the upper reservoir and you get a plant that can generate at maximum output for several hours before the upper pond drains.
That is not a long time in absolute terms. It is more than long enough to cover an evening peak. The daily cycle at Dinorwig looks something like this: pump overnight when the grid has surplus wind, generate through the morning peak, sit ready through the day, generate again through the evening peak, pump overnight again. Six turbines, one mountain, endlessly recycled water.
The plant does not consume water. The same molecules go up and down the shaft for decades. Evaporation and the occasional top-up from rainfall are the only inputs.
Why grids need machines like this
A modern electricity system is a strange thing. Supply must equal demand at every instant, and the consequences of getting it wrong are severe. Too much supply and the frequency climbs, damaging equipment. Too little and it falls, tripping protection relays and cascading into blackouts. The margin for error is measured in fractions of a hertz.
The grid used to keep its balance by leaning on the spinning inertia of huge coal and nuclear generators. Their rotors, weighing hundreds of tonnes and turning at 3,000 rpm, resisted sudden changes in frequency by simple mechanical stubbornness. As those plants retire, the inertia goes with them, and grid operators are casting around for replacements.
Pumped hydro turbines provide the same service. The rotors at Dinorwig weigh hundreds of tonnes each, and they spin whether or not water is flowing through them. When the grid frequency dips, the machines resist the change automatically, buying the operator seconds to bring on more power. It is the sort of invisible service that only matters when it is missing.
The Welsh station in a longer story
Electricity has been won and lost on questions of speed since the beginning. Energy Daily has previously written about Tesla’s 1893 demonstration of alternating current at the Chicago World’s Fair, which settled the format of the grids that Dinorwig would one day plug into. And we have covered the 1901 gusher at Spindletop, which unleashed the cheap oil that later kept gas peakers competitive with hydro for half a century. Both stories are prologue to what is happening now inside Elidir Fawr.
The mountain is a machine that was built for the coal grid and inherited by the wind grid. Its job description has changed, but its physics have not. Water still falls. Turbines still spin. The kettles still go on at half past ten.
On a still evening in August, when the wind farms of the Irish Sea are becalmed and the sun has dropped behind the ridge, the phone rings inside the control room at Dinorwig. Somewhere above, a valve opens. Water that was lifted to the top of the mountain during the previous night begins its fall. Seconds later, more than a gigawatt is on the wires, running north to Manchester and south to Birmingham, and nobody watching television in either city notices a thing.
