Pumped hydro doesn’t have to be expensive: If it is, it might be in the wrong spot

Australian pumped hydro energy storage (PHES) project proposals tend not to be located at premium sites, which translates to higher cost projects.

Australia has 300 premium (Class AA) pumped hydro sites listed in the global pumped hydro atlas in the size range 15-5000 GWh. For perspective, 5000 GWh is the effective storage in 100 million EV batteries.

The key parameters for low-cost premium PHES are explained in our new review paper: 

– Large head (>500m) is the most important geographical factor. Using a site with the same volume of water but with double the head will double the energy storage (GWh) and double the storage power (GW) but much less than double the cost.

– Large W/R ratio (>10). A site with double the ratio of the volume of stored water (W) to the volume of rock (R) needed to construct the reservoir walls translates to half-cost reservoirs.

– Large slope (>10%). Tunnels are expensive. Large ratio of head-to-spacing between the reservoirs (= slope) is strongly favoured. An aqueduct or low-pressure pipeline can often shorten the pressure tunnel. 

– Large scale (>40 GWh): a single 80 GWh PHES system is much more cost-effective than five 16 GWh systems. It’s cheap to make larger reservoirs to store more energy, by scooping up more rock to make higher dam walls.

– Long duration (>100 hours): a 160-hour PHES system (eg 80 GWh @ 0.5 GW) has ten times lower water and electron flow in the tunnel, pump/turbine and transmission compared with a 16-hour (80 GWh @ 5 GW) version using the same reservoirs, allowing much lower costs. A low-flow pressure pipeline could avoid tunnelling, and lower voltage transmission could be used. 

PHES is ten times cheaper than batteries for large-scale energy storage (GWh) while batteries are cheaper for high power (GW). Long duration (e.g. 160-hour) premium PHES as a hybrid with four hour batteries (located at grid strong points) combines large-energy, high-power, high-peak-period selling price and negative-cost recharging, by allowing storage to trickle-recharge storage as explained here.

Snowy 2.0 is Class AA with storage of 160 hours. Its long tunnel is offset by not having to build reservoirs. From 2028, it will provide 85% of NEM energy storage (350 GWh) at 10X lower capital cost ($34/kWh) than equivalent batteries and with five times longer lifetime. 

Selected premium (Class AA) sites are shown in the Table together with URLs for visualisation. All are outside national parks (except Snowy 2.0), and all have good access to local workforce, transmission, roads and water.

For perspective, Australia needs about 1000 GWh of additional storage to fully decarbonise via electric-driven transport, heating, industry, metals, chemicals, and synthetic aviation fuel. 

However, most proposed Australian pumped hydro projects have combinations of small head (200-400m), small W/R (1-5), small slope (3-8%), small scale (8-18 GWh) and short duration (8-18 hours). Costs derived from such sites are poor guides to long duration Class AA sites.

Andrew Blakers is professor of engineering, Australian National University