Amidst the rapid growth and constant news cycle around battery storage, it can be easy to forget that we’ve been using pumped hydro for electricity storage for decades. In terms of energy capacity it makes up over 95% of storage today. Even with the rapid growth of batteries, some forecasts suggest that pumped hydro is still likely to account for around 50% of the total by 2030 (e.g.: IRENA).
This isn’t a technology whose development seems consigned purely to history. New pumped hydro capacity continues to be built: from about 160 GW today, almost 240 GW could exist by 2030 (source: IHA).
But why and where might new pumped hydro schemes be built?
Batteries have some clear advantages: mass, cross-sector manufacturing leading to rapid unit cost reduction, fast installation at a huge range of locations and scales, high round-trip efficiency and instant response times (to name but four). By contrast, pumped hydro schemes need site specific design and construction, involve lengthy permitting processes and long construction times and will generally benefit from the best “levelized cost” economics only when built very big.
On the other hand, pumped hydro schemes are built to last for many decades. Batteries – at least the Li-ion ones which currently dominate – are often warrantied for just a single decade (and that within strict rules as to permitted cycling behaviour and frequency). Some big pumped hydro schemes can store total energies with discharge durations of several days, whereas “long duration” battery applications these days are considered to be 4-hour-plus; where the “plus” may – rarely – mean up to 6-8 hours. Indeed many batteries, most in countries like the UK, are currently built to provide particular energy at particular rates (power outputs), but not for very long (often < 1 hour).
To fund their large, single-project capex requirements, pumped storage schemes need policies and financial cases based around long-term returns, a strategic view of power system infrastructure and perhaps other socioeconomic considerations (they employ lots of people to build). That’s likely to draw a very different pool of capital and investors to those building business cases for battery projects. In the latter, assets need to pay back much more quickly and investors have a wider range of project types and structures available; suiting a range of different risk appetites.
So, in practice, pumped hydro and battery deployment are two very, very different undertakings. They will fit into future power systems with very different purposes and economic cases.
Pumped hydro schemes will likely continue to provide an option for the task of large-scale, centralised, bulk energy shifting over long timescales. They could compete with alternatives like hydrogen production, distribution and long-term storage more than with batteries. Of course the latter will continue to eat into the market for all types of big, centralised storage; through the advantages of a more distributed, location-specific, application-targeted and quick-to-market range of energy storage solutions and fast-reacting grid-services needs.
Assuming business and investment cases can be made – a discussion for another day! – where will pumped hydro growth occur?
Below I’ve compiled a few current examples to illustrate the possibilities, grouping into these five categories:
- Greenfield sites
- Modified or Repowered sites
- Brownfield sites
- Co-located sites
- Coastal sites (seawater)
No new pumped hydro schemes have been developed in Britain since 1974, when work began at the Dinorwig scheme in Wales. However, SSE has already been granted permission for a 600MW-capacity scheme in 2013, and has sought approval to increase that scheme to 1500MW. This project, Coire Glas in Scotland, would have a storage capacity of up to 30GWh, more than doubling the existing pumped storage capacity of the whole UK (currently 24GWh).
At a full 600MW it will take 50 hours to release this stored energy, whereas at a full 1500MW it will take 20 hours. The scheme is designed to “start generating electricity in less than 15 seconds when in a spinning cycle; and within 2 minutes from rest”.
Elsewhere, as with most aspects of energy these days, the biggest projects are usually happening in China.
So no surprise that the “world’s largest pumped storage power plant”, when it is completed in 2021, will be there. The 3600-MW Fengning plant, in Hebei Province in China, will be equipped with twelve 300-MW pump-turbine units, housed in an underground cavern. State Grid Xinyuan Co. Ltd. is developing the facility.
It’s a good example of long development timescale: preliminary preparation began in 1996! Site selection was completed in 2001, feasibility approved in 2010 and officially issued in 2012. Construction began in May 2013 and 1800MW is due to be commissioned this year. The total cost of the project is reported to be US$1.87 billion.
So far I’ve not seen any information on the duration potential (i.e. total GWh storage).
It’s an example of updated pumped hydro technology, using variable speed drive turbines to provide “up to 10% more power production in turbine mode” and “a reduction of up to 15% in electricity consumption” in pumping mode.
Repowered and Modified Existing Hydropower Sites
Another huge new pumped storage proposal is Snowy 2.0 in the Southern Alps of Australia, currently at the planning and approvals stage.
The “Snowy Mountains Scheme” already exists there: it’s huge set of infrastructure which integrates water storage, regulation and management – counteracting the impacts of droughts within the Murray Darling Basin – with hydro-electric power.
Nine major hydropower stations total up to 4100 MW of capacity and produce on average 4500 GWh of electricity each year, 32% of all the renewable energy available to the eastern mainland grid of Australia.
Snowy 2.0 is a major expansion of the existing Snowy Scheme, adding new pumped-hydro capability through an underground network of tunnels and a new underground power station. Up to 2000 MW of capacity will be added with about 175 hours of energy storage (350 GWh). (Up to 8000 MW of additional pumped storage are envisaged as potential from future schemes).
The first energy generation from Snowy 2.0 is expected in 2024 and the scheme is estimated to cost between $3.8 – $4.5 billion. (Future projects would utilise previously-built access routes and support infrastructure, so could cost less).
On a smaller scale is the recently commissioned Frades II pumped-storage project built in Portugal, by EDP Group (the main electricity provider).
It’s located in north-western Portugal and is actually the second repowering of the Venda Nova scheme, consisting of two dams built in the 1950s and 1960s.
In 1996, the original scheme was repowered with the Frades I plant, also a pumped-storage facility, with a new hydraulic circuit and a powerhouse equipped with two pump-turbines with an output of 97 MW each. That began operation in 2005.
Frades II construction began in 2010. This project too uses two variable-speed pump-turbines. In turbine mode, each unit can reach full-load of 390 MW in about 80 seconds. When pumping, the load varies between 300 MW and 390 MW. (For those interested in the technical and construction description of the project, the link I gave above is quite detailed).
Suitable greenfield sites are hard to develop, facing big environmental impact, public and planning barriers. So in recent years there’s been a lot of interest in identifying other sites – those where pumped hydro can provide a means to make use of site previously used for other industrial activities: in particular mining or quarrying.
Here’s one proposal to make use of the site of an old gold mine: in Bendigo, Australia.
The idea is to create a pumped hydro system with a capacity of up to 30 MW and up to six hours (180 MWh) of energy storage, taking advantage of around 5,000 shafts, as deep as 1400m and many hundreds of kilometres in total length, in a region mined for over 150 years. Most are disused but require management to avoid groundwater bringing to the surface a variety of nasty pollutants. So a key ancillary benefit of the proposed scheme is to tie in with this already-existing water management need.
The town already has a high installed base of solar PV. So water would be pumped to higher levels underground during the day powered by these local renewable energy sources then, at night and during late afternoon and early evening peak demand periods, allowed to flow back down to the lower levels driving turbines. A pre-feasibility report by Arup, suggests a round-trip efficiency of approximately 70%.
Here in the UK, the Glyn Rhonwy Pumped Hydro project in north Wales seeks to use abandoned slate quarries at different heights as upper and lower water reservoirs. As with most pumped hydro schemes, much of the development would be sited underground, with an tunnel connecting the two reservoirs via an underground turbine house. The grid connection too will be underground.
It’s worth noting that schemes like these require permits for initial filling, in this case from a nearby lake, along with ongoing access to water on a smaller scale (to counteract evaporation losses).
The Glyn Rhonwy scheme is designed with a capacity of 99.9 MW and 700 MWh storage capacity. It can go from start to full power in 12 seconds and developers claim a respectable 81% round-trip efficiency.
It’s worth noting that these brownfield examples are not much beyond the scale of battery projects now (and certainly not in the near future). Which means the advantages of brownfield development – both in terms of cost and local planning advantages – will need to be a key focus in making the case for development.
Co-location with Solar or Wind Power Generation
It’s rare that the topography helpful for pumped storage (hilly) coincides with that which lends itself to easy solar or wind power projects (flat). However there are exceptions, such as at the Kidston Renewable Energy Hub, in northern Queensland. Following on from those previous examples, this is a brownfield site too: another former goldmine.
Two former mining pits will form the upper and lower reservoirs for the pumped storage, which will be able to deliver 250 MW and 2,000 MWh (8 hours). It’ll be co-located and integrated with a 270 MW solar PV project. Already at the Kidston site is 50 MW of solar and 150 MW of wind power is at feasibility stage.
K2H, the pumped hydro project is estimated at $330 million cost. If approved, construction is due to commence in 2019 and should be completed by 2021.
Coastal: Seawater Pumped Hydro
Finally, when inland sites are hard to find or water is scarce (such as in arid regions), another option is to use seawater, with the sea itself as the lower reservoir. There has only been one seawater-pumped hydro plant in the world, the 30 MW Yanbaru Power Station in Japan. That plant ran from 1999 until decommissioned in 2016.
A problem with seawater systems is that they face the issue of corrosion, operating in marine environments, and ‘biofouling’ (which can block water in-flows). These challenges need managing through material selection and through biofouling prevention of various types (mechanisms already used in desalination plants for example). The economic result is both increased capital cost and higher O&M cost over the life of the plant. The lifetime of the plant is also likely to be shorter than freshwater pumped storage facilities, potentially around 30 years.
The Cultana Pumped Hydroelectric Scheme is one new seawater energy storage plant being planned in South Australia. It aims to have 225MW of generation capacity for up to 8 hours (1770 MWh).
It’s at an early-stage engineering design stage, after which the hope is for a 2-year build time and operation as soon as the summer of 2020/21 (though perhaps more likely around 2023).
Are there enough sites?
Some countries obviously have greater potential to expand pumped storage than others. It won’t make sense for everyone!
Traditionally, pumped storage has required dams located in river valleys. That brings big challenges around permitting, planning and environmental impacts that are likely to be insurmountable in many cases. Hence there’s such interest in repowered or brownfield sites, particularly when pumped storage can tie in with other, pre-existing water management requirements. Many such sites are “off-river”, simply requiring the existence or easy construction of two reservoirs at different heights. They do tend to offer lower durations though (typically from 5-25 hours, rather than the multi-day possibilities from giant conventional schemes).
One study in Australia identified 22,000 sites, all outside national parks and urban areas and each with anywhere between 1 – 300 GWh of storage. The same study concluded that just 0.1% of these sites would be needed for Australia to support a 100% renewable electricity system. Similar studies elsewhere have identified large theoretical potential: for example 15GW of “low planning risk” sites in the UK.
Of course there is often a wide gulf between theoretical potential and actual deployment! It’s a gulf filled with the choppy waters of competing solutions, financial viability, local acceptance and policy support.
Nevertheless, given the increasing importance of energy storage in power markets with growing capacities of solar and wind power, pumped storage may still prove to be a solution whose relevance doesn’t just lie in the past.
(First published via LinkedIN, here)