Groundwater refilled the tunnels of a South Australian zinc mine for 5 years after the pumps came out, and that flooded rock was exactly what a compressed air plant needed before the project was dropped

The decline at the Angas mine runs down into the hill on a gentle slope, wide enough for a truck.

Somewhere past 790 feet it disappears under still water.

The pumps came out at the end of 2013, and groundwater has been climbing back ever since.

Grid connection approval landed in September 2026 and operation is targeted for the end of the decade, so the claim about how it works is still a claim.

Nobody was planning to drain it.

Then a Canadian developer looked at that flooded rock and saw a place to keep electricity as compressed air.

Not despite the water. The water was the whole point.

Why the water has to stay in the tunnel

Compressed air wants to expand. Seal it in a cavern and the pressure drops as you draw it down, so the turbine at the far end is chasing a moving target all the way to empty.

The fix is to give the cavern a second fluid that will not compress, and water is the one already sitting in a closed mine.

A shaft of water rises from the cavern to a pond at the surface. Air pushed down displaces water upward. Air drawn off lets the water fall back.

The weight of that column holds the cavern at constant pressure for the whole cycle.

It also means every cubic foot of air that went in can come back out, with no cushion left behind to stop a vacuum forming.

Which is why a flooded mine is not a ruined site. It is half the machine already built.

What the heat does while it waits

Squeezing air makes it hot. Releasing it makes it cold, cold enough to damage the equipment it passes through.

The two old plants in the world solved that with a burner. Air comes up, natural gas is lit, the warmed air turns the turbine.

The newer approach strips the compression heat out of the air on the way down and parks it in an insulated store at the surface.

Hours later the same heat goes back into the air on its way to the turbine, and no fuel burns at any point.

Round trip efficiency lands somewhere near 60 to 66 percent, well under a lithium battery and well over nothing.

The trade is duration. Eight hours out of a hole in the ground costs far less than eight hours out of a cell stack.

What was actually approved near Adelaide

The site sits about 37 miles southeast of the city and produced zinc, lead, gold and silver until it closed.

The proposal was deliberately small. Five megawatts of output and 10 megawatt hours of storage, which is roughly two hours at full power.

The state put in 3 million dollars from its Renewable Technology Fund. The federal agency committed 6 million more.

Total cost was quoted at 30 million dollars, and later listed at 33.

Development approval came through in July 2019, with the storage cavern planned at around 790 feet and about 40 construction jobs attached.

Every announcement called it Australia’s first.

What the record says happened next

It was never built.

In January 2021 the mine owner confirmed the developer had walked away, pointing at grid usage charges, capital costs and pandemic disruption.

The federal agency’s project page now carries one line about it. The work has been discontinued, and the amount actually paid out was zero dollars.

So the grants were real, the approval was real, the flooded tunnels are still there, and nothing was ever connected to anything.

The comparison numbers that travel with this technology deserve the same care. The developer claims up to ten times less water and twenty times less land than pumped hydro, against a dam 400 feet high.

Only one plant of this design runs anywhere on earth, a 1.75 megawatt unit in Ontario finished in 2019. Everything larger is still on paper.

Which is worth holding next to Queensland, where 780 megawatts of battery storage is being built on a schedule nobody is arguing about.

Where the idea went instead

Broken Hill got it. The town sits at the end of a single transmission line and goes dark when that line fails.

The plan there is 200 megawatts and 1,600 megawatt hours, eight hours of output from a cavern roughly 2,000 feet down in hard gneiss.

That is 160 times the energy of the South Australian design, and the price has moved to about 652 million dollars.

The cavern is not a found void either. It gets excavated, sealed behind a concrete bulkhead and monitored, with the old mine supplying access and surface ground rather than the space itself.

Grid connection approval landed in September 2026 and operation is targeted for the end of the decade, so the claim about how it works is still a claim.

What the Angas file proves is narrower and more useful. A flooded mine and a surface heat store will not carry a project on their own, and the grant is the easy part.