Pumped hydro: Storage solution for a renewable energy future

Print Friendly

Of the various forms of renewable energy storage, pumped-storage hydropower is the most cost effective. Compressed air storage is another technology that is generally about 10 times smaller in capacity than hydro, much more expensive and requires large sealed underground caverns, which are rare. Although battery systems are, at most, 1 per cent of the capacity of typical pumped hydro and an order of magnitude more expensive, many dispersed batteries in residences, commercial premises and electric vehicles may also make a viable contribution to energy storage in future.

The main advantages of pumped hydro are relatively low levelised cost of energy (LCOE), large capacity (can be more than 1GW) and almost instantaneous dispatchability (response); the 1.7GW Dinorwig plant in Wales can ramp up from zero to maximum power in less than 60 seconds.

Hydro energy is generated by two-way turbines which can act as pumps and generators and is calculated by the formula: E = mghe, where m = mass of water, g = force of gravity and h = head (height), e = efficiency factor

Efficiency factors are applied twice – when water is pumped up and when it flows back, giving a round-trip efficiency of approximately 76 per cent, made up of pipe friction loss (6 per cent) and turbine efficiency loss (18 per cent). Pipe friction increases as diameter is reduced and distance increased. To keep friction losses low, short large-diameter pipes are required.

A recent Western Australian study into 100 per cent renewable energy for the SWIS electricity grid proposes solar and wind-powered scenarios with ‘over-built’ 10,800MW capacity to supply 5,500MW peak demand. Short-term backup is provided by pumped-storage hydro and molten salt storage. Longer duration backup is by biomass gasification and biomass co-firing of molten salt storage in concentrated solar thermal plants.

Stand-alone hydropower is not possible in WA and SA because flow is limited. There is limited potential for pump-back hydro on the existing small water supply dams. Pumped ocean storage hydro holds most potential for large-scale dispatchable power. Sites need to be within 2 km of the coast, >100 m in elevation and >100 ha in area, many of which have high conservation values. Nevertheless there a limited number of suitable sites within 100km north of Geraldton and east of Albany in WA, and the Fleurieu and Eyre Peninsulas and Port Augusta in SA.

A cliff-top salt water storage in WA would be located on land of 100-140m elevation about 1km from the shoreline. To provide 1500MW of capacity, 15-20 headrace tunnels or pipes about 5m in diameter feed hydro turbines, with intake/discharge into the ocean. To minimise environmental impacts and maximise cost effectiveness, two 200ha clay/membrane sealed cliff-top ponds with 120-140m head would be optimal; there are several sites fitting these criteria. Each power station would provide 700-800MW of fast response backup power for up to six hours and would stabilise the output of future wind and wave generators in the vicinity. Such as system would cost less than $5 billion.

Increasing the pond area fourfold would give a full day’s energy supply but would have to be weighed up against the environmental and cost impacts of the larger ponds. To put size in perspective, some natural coastal inlets on the SW coast are >2000ha.

From the tables below, we estimated the capital cost of pumped ocean storage in WA to be $2500/kW capacity, assuming a 10 per cent discount rate over 40 years, 7 per cent additional cost for operating and maintaining seawater systems and a life of 80 years. Energy cost would be about 8.6c/kWh (from IPCC table below). Given the excess wind energy stored would otherwise be wasted, it would cost only a few cents/kWh and it should be possible to supply low cost dispatchable / ‘peaking’ hydro energy at an LCOE of less than 14c/kWh. Currently the only commercial pumped ocean storage hydro power plant is a 30MW plant in Okinawa, Japan. A 480MW plant is planned for Glinsk, Ireland and another, 300MW in Hawaii.

It is regrettable that Australian state governments, faced with the huge dual problems of energy insecurity/high gas prices and global warming have not yet produced in-depth costed reports identifying optimal renewable energy electricity systems, including the transmission lines and smart distribution grids required to enable them. SEN hopes that its report will give them a ‘nudge along’ to do so.

Screen Shot 2013-04-16 at 10.56.27 AMScreen Shot 2013-04-16 at 10.58.36 AMBen Rose is a member of Sustainable Energy Now and co-authored the report ‘100% Renewable Energy
on the SWIS Grid 2029’

RenewEconomy Free Daily Newsletter

Share this:

  • adam

    Thanks for this.

  • Terry Wall

    Interesting article..

    What about smaller scale town size or community size plants. Big is not always better as is evidenced by PV on houses. Tanks even, using fresh water.

    What about remote locations where such technology would allow communities to avoid the huge cost of power lines and losses. Has to be a no brainer really..

    • Ben Rose

      Terry, I used the E = ‘mghe’ (dividing by 3.6 to convert MJ to kWh) for a hypothetical 100,000 L tank on a 10 m stand or hill). That’s quite a big tank holding 100 tonnes of water. I came up with about 2 kWh of storage or about 1/5th of an average household’s daily consumption. That would be an expensive project just for that much electrical energy storage; I’ve only seen one such tank on a remote community and it was used to cool the hot artesian water they were using. To achieve even this much, the tank would have to be very close to the pump/turbine as friction rapidly consumes the energy of water flowing through the small (<150mm) pipe you'd be using.

      So I think for small scale dispersed applications, Li-ion batteries would be far cheaper; about $8000 for a 20 kWh battery pack like those used for EV's.

      Pumped hydro is really only cheap on a large scale.

  • James Fisher

    Hi Ben,

    Thanks for the analysis.

    You quote $2500/kW and I assume this is your turbine nameplate capacity but what size do you assume for your reservoir? If the reservoir is big enough to run the system for 4 hours it will cost a lot more than a size for 10 minutes. The LCOE will account for the storage size but can you specify the kW/hr capacity as well?

    • Ben Rose

      James, re your point about reservoir size, yes this would increase the cost considerably. I don’t know how many hours of storage NREL’s diagram assumes. But if it were 6 hours (a likely figure) and we wanted to increase it fourfold to 24 hours and wall height stays at 10 m, pond area would have to be increased 4 fold, while the rest of the system would be unchanged. To do this the wall length would only have to double as area increases to the square of perimeter (assuming a circular pond). Sealing costs would be proportional to area. So in this case I’d say quadrupling storage would likely increase cost by somewhat less than 3*$420 per KW; say 1000/2500 or 40%.

    • David

      James, Pumped Hydro is virtually always sized for 6 to 8 hours of generation at rated capacity because all of the plants to date have been designed for use with baseload power. For this traditional application there is no point in having more than 6 to 8 hours of storage because there is always surplus power (from coal or nuclear generation) to “recharge” the storage the next night. Renewables require a paradigm shift and it would be better to have much longer duration storage because wind and solar lulls can last many days.
      I agree with Ben that pumped hydro is our best option for grid scale storage but I would say go bigger on the reservoirs (or relatively smaller on the generation). As Ben’s reply suggests, the bigger the reservoir, the cheaper the energy storage in $/kWh.

  • David Clarke

    I have looked into possible sites in South Australia. (

    In addition to the (mentioned) Fleurieu Peninsula, the north coast of Kangaroo Island (both on the Dudley Peninsula and ‘the mainland’) also has potential sites, and is not a long way from Adelaide.

    • David

      I have also looked at possible sites for pumped hydro in SA. I would suggest a much larger scheme on the Simmens Plateau feeding into the upper Spencer Gulf. It would require two large dams (the combined wall comparable in size with the Aswan High dam in Egypt) giving a single reservoir of volume 1100 Gl and an active storage capacity of 347 GWh with a working head of 189m. The problem would be increased salinity in the upper Spencer Gulf.

  • sunevaenergy

    Good info ,looking for sites!

  • Scottish Scientist

    World’s biggest-ever pumped-storage hydro-scheme, for Scotland?

    The map shows how and where the biggest-ever pumped-storage hydro-scheme could be built – Strathdearn in the Scottish Highlands.

    The scheme requires a massive dam about 300 metres high and 2,000 metres long to impound about 4.4 billion metres-cubed of water in the upper glen of the River Findhorn. The surface elevation of the reservoir so impounded would be as much as 650 metres when full and the surface area would be as much as 40 square-kilometres.

    The maximum potential energy which could be stored by such a scheme is colossal – about 6800 Gigawatt-hours – or 280 Gigawatt-days – enough capacity to balance and back-up the intermittent renewable energy generators such as wind and solar power for the whole of Europe!