Our website uses cookies to improve your user experience. If you continue browsing, we assume that you consent to our use of cookies. More information can be found in our Cookies and Privacy Policy. 

World-first in grid-scale pumped heat energy storage places UK at forefront of energy storage R&D, team claims

Pumped heat energy storage (PHES) shuffles heat between two tanks containing mineral gravel by means of a working gas, generally an inert gas such as argon. In storage mode, the argon is pressurised to around 12 bar, which heats it up to 500°C. The hot gas enters the top of one of the tanks (designated hot), and flows down slowly at about 3m/s, heating the particulate in the tank and itself cooling down. At the bottom of the tank, the gas is still at 12 bar but now at ambient temperature. At this point, it is expanded back to ambient pressure which cools it to -106°C. It then enters the second tank, cooling the particulate and itself warming up, exiting the tank at the top at ambient temperature and pressure.

To recover energy, the process is reversed with ambient argon entering the cold tank, being cooled and becoming pressurised. It then enters the hot tank, where it is warmed to 500°C but remains at the same pressure. It returns to ambient pressure in the expander, which drives a generator. The round-trip AC-to-AC efficiency is claimed to be 75 to 80 per cent, with the energy to be stored driving the compressor in the charging phase.

The Newcastle team has assembled and commissioned the system at the Sir Joseph Swan Centre for Energy Research. It is rated at 150kW and is capable of storing up to 600kWh of electricity.



Originally developed as part of a distribution scale energy storage project funded by the Energy Technologies Institute, which aims to test and de-risk new energy-related technology, the system is designed to store energy generated by renewable sources to buffer their inherent intermittency. The team at the Swan Centre has operated the system in both expansion and compression modes and claims it can switch between discharge and charge in a few milliseconds, according to Andrew Smallbone, co-director of the National Facility for PHES, who led the project.

“Pumped Heat Energy Storage or Pumped Thermal Energy Storage is cheap and is compatible with the technical and scale-up challenges of grid-scale energy storage,” said Prof Tony Rosskilly, director of the Swan Centre. “Given the thermal power cycle’s enormous potential, there has been a tremendous amount of research and commercial interest in PHES technology over the last ten years, however until now nobody has managed to get as far as to demonstrate a real-world working system. What is exciting is that the UK is the first to do it and as such, is now leading the world in what looks like a highly disruptive and cost-effective technology which can balance renewable energy supply and demand.”

Smallbone’s team has not yet achieved the full potential efficiency of the system, but he said that the 60- 65 per cent efficiency it has obtained is consistent with the original target design specification and is high enough to place the technology as the lowest-cost and most flexible grid-scale energy storage technology currently available.

“Additionally, these tests also indicated there is significant opportunity for further improvements through design enhancements and operational optimisation. This will now continue over the next few months,” he said.

Visit the UK’s dedicated jobsite for engineering professionals. Each month, we’ll bring you hundreds of the latest roles from across the industry.

Oops! What happened with the cycle efficiency? Did an engineer goof on the figures, or are we losing heat somewhere? Surely, they will surmount the missing 10% efficiency and move it up the scale, perhaps by making a larger unit next time?

This project does not appear to be focused on a peak efficiency as much as investigating viability in a grid attached (11KV) storage application by weighing factors such as cost against various other metrics such as efficiency, technology maturity of available components, operational characteristics etc. The theoretical basis underlying the project was data from a number of preexisting models which vary in detail but between them define a general theoretical range of efficiency. There is a paper (see below for details) which gives some details of the pre-existing models and some possible design variants in terms of target efficiency vs implementation cost and also mentions effects of operational characteristics e.g. rate of discharge.

Levelised Cost of Storage for Pumped Heat Energy Storage in comparison with other energy storage technologies. Andrew Smallbonea, Verena Jülchb, Robin Wardlea, Anthony Paul Roskillya Sir Joseph Swan Centre for Energy Research, Newcastle University, Newcastle upon Tyne NE1 7RU, UK b Fraunhofer Insitute for Solar Energy Systems ISE, Freiburg, Germany

This is a “first operational unit” and as such, is still not fully optimised. I suspect that a larger unit would have higher cycle efficiency, and that future units will be further optimised to more closely approach theoretical efficiency levels.

Eureka! the gas becomes pressurised as it cools down. Then it can be expanded to generate electricity. I would expect the efficiency to be higher.

What is the expected life of the particulate filling in the tanks? As the particles expand and contract with the temperature cycles the particles will grind against each other resulting in a change to the particle size distribution. This could have a marked impact on the ability of the gas to flow through the particle bed.

Hi, I was the inventor and CTO of Isentropic Ltd when we designed this system. In response to the above two questions. I should add the caveat that I have not been involved with the project since the demise of Isentropic Ltd in 2016, the following notes being based on the design as intended:

Cycle efficiency. The efficiency of a full scale system was estimated to be around 74%. This assumes that ancillary systems are designed to a high standard to match the machine, this is not practicable for a first of kind test article. The test machine is also still sub-scale and the 74% assumed a MW scale machine. 60-65% for the machine as built actually exceeds my expectations for that test article.

The life of the particulate filling is an interesting question and one that excercised me considerably during development. We used magnetite with a grain size of 2-3mm in a layered structure within the tanks. The layers serve two purposes; the first is that with controllable baffles and a peripheral flow bypass path, layers that are not being actively heated or cooled can be bypassed reducing pressure loss and also controlling thermocline degradation (thermodynamic irreversibility) during successive cycles. The second is that each layer is quite thin (about 110mm) meaning that the hydrostatic load on grains is reduced. This is vital to avoid thermal ratcheting with each cycle that damages grains and can create very high stresses.

Magnetite is fairly soft and very dusty indeed in its untreated form so we also fired the graded material to harden the surface. It should have a very long life but only cyclic testing would confirm that. The system is fairly resistant to getting dust into the gas circuit as the velocity of gas through the storage layers is very low and filters are placed across the entire store area at entry and exit.

I am not under any non-disclosure agreements with respect to this project and can therefore answer any questions of interest without restriction.

I hope this provides adequate answers to your questions. I still regard the pumped heat concept as the only viable large scale storage technology anywhere in sight at present with respect to cost, life and efficiency and regard the near abandonment of this project by the UK as a seriously wasted opportunity. It has taken a ridiculous amount of time to get to this stage (I concieved it in the late ’90s and started proper development work on it in 2004) and I doubt very much whether it will be developed to a worthwhile level within the UK. Other well motivated groups are now engaging with the concept around the world so it is likely to get there in the end in spite of the UK’s well known resistance to innovation.

Hello Jonathan, as you have volunteered to answer questions about this – thanks – could you provide a concise explanation of how by cooling the ambient argon it ‘becomes pressurised’?

Obviously I didn’t write the press release. This is gobbledegook! Obviously the gas density increses as it cools within the store but also obviously, this does not pressurise it. The charge cycle is: Adiabatically compress from system datum temperature and pressure (to about 500 deg c, 12 bar, arogn is the working fluid). Pass to hot store and cool gas within the store while leaving the heat in the store. Exit hot store, still at 12 bar but at close to system datum temperature (there is an exit heat exchanger to ensure that datum temperature is maintained). Expand adiabatically to original pressure, this drops the temperature to about -160 deg c (not -106 as in the press release). Pass to cold store cooling the store and warming the gas back to system datum temperature.

The discharge cycle is exactly as above but in reverse, substitute wrming for cooling, expansion for compression. As the engine efficiency of a reversible thermal cycle is the inverse of the same cycle used as a heat pump, the limit efficiency is the product of the two. Anything multiplied by its inverse=1, hence Carnot restrictions cancel out in this reversible system.

Thank you for this clarification. The description didn’t come from the press release, but from the info page on the process from the Joseph Swan Centre website, which could obviously do with some editing. It’s been far too long since I studied thermodynamics, or I wouldn’t have incorporated the error.

As does my 30+ years old grasp of thermodynamics. The late Dr Frank Palmer would be thoroughly ashamed of me.

Thank you for your useful interjection. And I can only say how sorry I am to hear of the lack of support you had; sometimes we see behaviour in our chickens when they all race over to to where one has found something – ignoring their necessary proper feed ;-{{. I think now people are starting to appreciate the importance of energy storage (and sensible thermal). I think you have well answered my question about flow management We were looking at using similar structures (but for different purposes) and using standard sized (steel) ball bearings – which we thought might be able the better to withstand cyclic thermal shocks – is that something you considered? I hope your expertise is appreciated and that you are brought in to help develop the technology.

Thank you for the intelligent comment. Yes, gravel was chosen for low cost but this actually turned out to be a bit of a false god as the dominant cost is that of the containing pressure vessels. Given the trouble that we had with heat treating the magnetite and grading it I would now use steel shot or similar as the final cost would be lower and it is much cleaner.

The efficiency of a system like this is not related to heat engine efficiency. Before this makes anyone splutter let me explain:

This is a stringing together of four highly reversible processes, two, fast, adiabatic compression/expansion, two extremely slow isobaric heat transfer processes. The engine cycle has a Carnot ratio just like any other heat to work cycle, however, the heat pumping cycle used during charge is the exact same cycle with a thermal coefficient of performance equal to the inverse of the Carnot ratio of the engine cycle. In a round-trip operation these cancel leaving the system efficiency, in this respect, independent of any Carnot limits. This is the key to the whole concept.

Real losses such as heat leakage, pressure loss and mechanical are present and these are proportionately much worse with a low powered system than a high power system of similar sizes. A higher temperature span system will be higher powered and so the real world losses reduce with higher temperature spans. +500/-160 degrees c was selected to be (just) manageable with conventional materials. This choice did throw up some horribly tricky problems with respect to low friction piston sealing and valve design in an oil-free gas path machine and the solutions should, if properly exploited be valuable in their own right, although I will not hold my breath on that.

Positive displacment (reciprocating) machinery was developed as aerodynamic machines are simply not sufficiently reversible enough to deliver anything like a worthwhile efficiency as other projects are inevitably going to discover. The machinery in question is also far from conventional as valve pressure loss and timing precision requirements dictated a new approach. This led to one of the most difficult engineering problems that I have ever solved, namely the aerodynamic control of a very lightweight, fast moving, actively controlled screen valve (a form of reed valve). This should also be of intrest in efficient industrial compressor design, but agin, don’t hold your breath.

The key learning from this whole excercise has been the reality of trying to innovate in the UK. Sadly, we are not in a good place and there is a very good book in this still to be written as it will not get fixed until the problems are recognised.

Thank you for your kind comment. And the details about the thermodynamic cycles. I was looking at the field of high performance insulation (vacuum) and making it robust and affordable – so very similar indeed to the needs of your pressure vessel; and indeed it is a requirement for other industries, such as nuclear power. Unfortunately this sort of issue is dominated by the manufacturing technology – and this does not seem fashionable. I know that the status of innovation has improved, in government circles, – it is no longer “best practice” but, now, “new product”. Your journey shows that you had to consider (AND develop) technologies in many areas – which does not sit well with business plans as they stand. Historically people marshaled and developed (if not available) technologies ( and resources) to achieve innovation; Wedgwood and Paxton are good examples. However I do not think financial resourcers have bought into the radical innovation model – nor indeed manufacturing opportunities. PS I was looking at a German company who made vacuum insulation vessels for thermal storage; cheap materials but took a week to achieve vacuum!

Thanks for all the information Jonathan. CAES is certainly the best available storage for large scale, as in Germany and the USA schemes that have worked for many years. The small-scale version is fascinating as usually economies of scale are needed for prime-movers to be cost effective. You mention using piston engines for the expanders and compressors, surely turbo-chargers would be better suited: cheap and mass produced and suitable up to probably 1 MWe, were they ruled-out for some reason?

This isn’t CAES. It stores a property of the working fluid (heat), not the working fluid. With a form of transfer heat exchanger (not within the current suite of technologies of the Newcastle group) it can store indefinite amounts of energy in unpressurised form. This was something that I had in early stage concept before Isentropic folded, it did not get as far as a patent application.

Turbo- machinery introduces an unacceptable level of irreversible thermodynamic losses I thought I discussed that in an earlier post?). This is a storage system so every irreversibility bites twice, once on charge, once on discharge.

The machine as built has a uniflow construction with valves that use the entire bore area for both inlet and outlet. They are finely divided (thousands of small ports) and opened/closed with a matching perforated screen of low mass (typically 0.2-0.25mm thick Inconel) and high flexibility (better sealing).

When closed, pressure locks the valve in place and an opening force is applied in the plane of the valve. When cycle pressure equalises the friction locking the valve in place disappears and the valve snaps open very close, in cycle time, to the ideal time to open.

Closing it is a very difficult problem as it has to fly across the port and drop in precisely the right place, prior to locking, at ptecisely the right time. I did this with some rather involved micro-aerodynamics and a couple of mechanical details.

The screen of finely divided ports is very important. If a conventional valve is used then extensive swirl and tumble of flow occurs. This results in high heat exchange with cylinder walls and the head/piston crown which damages reversibility severely. The screen valve gives teh flow fine scale, evenly distributed turbulence with very little tumble/swirl. It also gives around 20% of the bore as an openb area reducing pressure losses. This was so effective that on presenting some cylinder indicator diagrams to technical specialists, they initially thought they were looking at theoretical cycle diagrams, not test results.

Sadly, the valves in the machine as built were the result of railroading a poor interim design through against my better wishes and I am very surprised to see the announcement since I did not expect this valve to operate in an expander mode other than at extremely low speeds. The proper valve was held back to “get something built” and I am not aware that proper knowledge of this valve and its subtlties exists within the current group. Happy to be proved wrong. We’ll see.

An interesting article – with some useful numbers in for discussion and appreciation. 600 units of electricity – equates to about 2×10^J or 20000 MJ. If specific heat is 2 J/cc/deg then 500 degrees equates to 100 J/cc or 100 MJ/cu m gives a volume of 100 cu m, for heat store; I thought this might give a scale to the picture – and the possible scale of a household unit (approx 100/60 cu m) I am surprised that the efficiency is so high, as gas turbine generators working at 1500 degrees only claim ~60% efficiency – but have not done the sums and looked at possible combined cycles. The use of gravel is, of course, cheap but, unless some flow channeling structures are embedded might cause problems when scaling up (as flow through packed beds are often diffusive) – and uniformity of “particles” and their arrangement might be an issue. I look forward to hearing how the technology goes and how upscaling (thermal powerplant) and downscaling (domestic), and associated issues, might be addressed – and their afford-ability.

There are currently at least grid scale thermal storage systems in commercial operation around the world, using molten salt as a storage medium. Typically over 100 MW with over 1 GWhrs of storage at over 560oC. See https://en.wikipedia.org/wiki/Crescent_Dunes_Solar_Energy_Project etc. Durability seems to have been proven while this is still in the experimental stage. So how does this technology compare in cost and efficiency?

I find it interesting that thermal management (including storage) is an important issue for solar and nuclear power – with many common flow and materials issues. The use of a noble or inert gas (Ar , He, N2) obviates many of the materials issues – such as corrosion – that beset many liquid working fluids – such as Sn, Pb or molten salts (but not sodium!). This obviation of high temperature corrosion means that these systems can work at much higher temperatures – producing process heat as well as more efficient generation of electricity. And I think Nitrogen and Argon would be well suited to using conventional gas turbine (off-shelf) machinery – possibly with reduced problems for blade materials! In addition the use of gas facilitates better management of heat (as the inventor says) – and does not have issues of the working fluid is a liquid and might freeze (for example Soviet Pb cooled submarines always had to be kept hot when undergoing maintenance – lest freezing occur – and this could be an issue with solar thermal using molten salt) – possibly why Sodium is one of the preferred coolant for some nuclear reactors (if temperature less than 880 degC). It is interesting to see that the less glamorous technologies (such as high temperature seals and valves or the manufacture of pressure vessels) are important in necessary in getting reliable and affordable technology. I do not know the costs of concentrated solar thermal – but I think the technologies that have been described, in the discussion, have got legs…

Taking a 11 gram minnow as descriptive of Swan’s pumped heat energy storage capacity of 600kWh and considering various capacities of pumped storage hydro schemes and calculating a ratio of these energy capacities with respect to Swan’s pumped heat energy storage capacity and using that ratio to scale up the weight of a minnow to find an appropriate weight of life-form to describe such an energy storage capacity …..

A “woman ice-skater” of energy storage. 2,400,000 kWh Red John Pumped Storage (proposed by ILI). At Loch Ness.

A “male rugby player” of energy storage. 6,300,000 kWh Foyers pumped storage hydro scheme. At Loch Ness.

A “strong man” of energy storage. 10,000,000 kWh. The largest of the UK legacy pumped storage hydro schemes, such as Cruachan or Dinorwig.

A “male polar bear” of energy storage. 30,000,000 kWh. Coire Glas Hydro Scheme. (Proposed by SSE)

A “blue whale” of energy storage. 6,800,000,000 kWh. The Loch Ness Monster of Energy Storage. 13 miles from Loch Ness.http://euanmearns.com/the-loch-ness-monster-of-energy-storage/

Strathdearn Pumped Storage Hydro Scheme. (Proposed by Scottish Scientist)https://scottishscientist.wordpress.com/2015/04/15/worlds-biggest-ever-pumped-storage-hydro-scheme-for-scotland/

No, it isn’t. It is a small scale technology demonstrator of a system that could be placed anywhere. Building a pathfinder system at a scale simlar to a large pumped hydro plant would be ridiculous for an unproven system.

Pumped heat has a much higher storage per unit land area than pumped hydro and is independent of special geography. Pumped hydro is an excellent storage technology where it can be applied. The pumped heat concept is designed to be distributed around a network, typically at sub-station level.

With comparisons between energy storage being mentioned I decided to check my figures ad search elsewhere ; the Wiki on Energy_density ( mentions pumped storage as 0.001 MJ/L) and gives a 500 degC storage unit as having 1MJ/L 2 (J/deg/cc x 500 = 1 kJ/cc ==> 1MJ/L) – so 1 unit (3.6MJ) requires 3.6 litres. Which gives a volume (for 600 Units) as 2160 litres (2.2cu m). This indicates that sensible heat storage is more efficient than pumped hydro.

However energy density is one thing(and it does indicate thermal storage could be viable for domestic usage) but one must consider, too, the cost models (capital) – and, in the case of large systems, the associated construction management capability in this country. The economic viability figures for seasonal and daily storage as 0.25 and 75 Euros/Unit — so a 600 Unit storage unit economic price of 150 and 450,000 (!) Euros respectively. (see https://iea-etsap.org/E-TechDS/PDF/E17IR%20ThEnergy%20Stor_AH_Jan2013_final_GSOK.pdf ) Costs for Thermal Energy Storage” Construction costs are about $52/Unit (for a two molten salt tank system) – with half being molten salt medium and the other half tank construction (So 600x$52 = $31200) And https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/545249/DELTA_EE_DECC_TES_Final__1_.pdf indicates UK Gov understanding and thinking; recognises need to understand heat losses for individual building heat storage – but seems to focus on what is available on market – water tanks and considers technology is mature!-?{ I appreciate this is a quick snapshot of numbers. I suspect that reducing the price of the technology would make it more relevant to domestic application. As high temperature energy is more valuable I would expect that the value would be higher than the water tank based analyses; however large or small systems would benefit from low cost manufacture.

Centaur Engineering Limited Registered Office: Mark Allen Group, St Jude's Church, Dulwich Road, London, SE24 0PB Registered in England No. 11569365

1hp Piston Type Air Compressor

Air Compressor, Silent Air Comppressor, Low Loise Air Compressor - GNGJ,http://www.gngjaircompressor.com/