Is ARES the solution to the energy storage problem?

Every so often an item appears in Blowout Week that’s worthy of further discussion, and Blowout Week 118 has one. It’s the article on ARES – Advanced Rail Energy Storage – a simple combination of three proven technologies – railroads, potential energy release and regenerative braking – which reportedly has a number of advantages over its numerous energy storage competitors:

* All it needs is a rail line, heavy rail cars with regenerative braking, and a hill.  It needs no reservoirs, pump houses,  penstocks, underground cavities, salt mines, submarine bladders or even water.

• Environmental impacts are usually minimal, energy efficiency, costs and ramp rates are reportedly comparable to pumped hydro, there are no limits on the number of charge/discharge cycles and no degradation with time.

• There’s no lack of ARES natural resources (hills) in many parts of the world. Storage capacity can be made as large or as small as needed in these areas.

The ARES concept has been tested at a pilot project in Tehachapi, California. No results are provided but some intriguing images are:

The ARES gravity train used at the Tehachapi, California, pilot plant

So let’s take a closer look at ARES:

First, sources of data. As is common with “first posts” I have used a number of basic data sources that tend to be repetitive, and rather than burden the text with lots of duplicate references I am listing all these sources below:

ARES Press Kit

Pahrump Rail Energy Storage Project

America’s First Commercial-Scale Rail Project Receives Approval

Advanced Rail Energy Storage uses heavy train cars to store power

Gravity train as energy storage

ARES system to put energy storage on the right track

ARES performance

Plus a potted overview from Utility Week:

Technologies vying for the bulk storage market

Now on with the show. The Blowout Week article describes how Advanced Rail Energy Storage LLC has just been granted a right-of-way lease by the US Bureau of Land Management for a 50MW rail storage project in Pahrump, Nevada. The approval came after an Environmental Assessment concluded that the project would have no significant impact (it disturbs only about 70 hectares). Details of the project are summarized on the list below and on the following project layout map:

      1.  A single 9.2 km long track with an elevation change of approximately 640m.
      2. Six 300-ton trains made up of rail cars with regenerative braking capability.
      3. 50MW peak output, 12.5MWh storage with the option of scaling up to 1 GW.
      4. Capital cost $55 million ($1,100/kW installed. I can find no information on costs/KWh.)
      5. “A lower life-cycle cost than batteries, a higher energy-to-power ratio than flywheels, and a greater efficiency and faster ramp-rate than pumped-storage.”
      6. An 8-month construction time, project life 30-40 years.
      7. “The project has all private financing, no government loans or grants.”

ARES Pahrump project layout map

The project will be grid-connected and designed to provide “fast response energy to assist the balancing of intermittent renewable energy [solar and wind]” and handle “momentary changes in demand”, which probably explains why the storage lasts for only 15 minutes at 50MW output. Long-term storage could, however, presumably be achieved simply by parking trains at the top of the hill and leaving them there.

And here’s what I understand to be an artist’s impression of what the rail yards for a 1GW system would look like:

A 1GWh(?) ARES facility

An entertaining video of how the rail cars are shunted around to compensate for changes in PV output as clouds pass over the panels is also available here. I can’t post a direct link to the video but it’s the image at the bottom. There are more images of interest above it.

Other technical features of the ARES system include:

Reactive Power Production – The shuttle-trains onboard Dual 3-Level Active Rectifier/Invertors are capable of supplying 25% of generated system power as reactive power for grid VAR support in full discharge mode and in excess of 100% of system power as reactive power while synchronized to the grid in standby.

Heavy Inertia – When in direct grid synchronization the ARES shuttle-trains provide beneficial heavy inertia — augmenting grid stability against the loss of heavy generating facilities and increasing reliance on solar energy.

High Efficiency Regulation – While providing Regulation-Up and Regulation-Down support to the ISO a separate dedicated pool of loaded ARES shuttle-trains are available to dispatch from mid-system elevation complying with ISO regulation commands without having to overcome the efficiency loss of operating on pre-stored energy. As such an ARES facility is able perform a round-trip regulation Reg-Up/Reg-Down command at over an 86% operating efficiency.

Variable Output at Constant Efficiency – Unlike CAES and pumped-storage hydro there is no loss of system pressure during discharge. ARES system efficiency is constant over the full range of discharge and power output.

It’s nice to know that ARES considered and apparently resolved its grid stability issues before starting construction, which is more than can be said for another project Energy Matters has been discussing recently.

And the System Ratings image below reportedly demonstrates the superiority of ARES over all other energy storage technologies except pumped hydro:

ARES system rating versus other energy storage system ratings.

But ARES has a major advantage over pumped hydro too. Pumped hydro needs favorable topography and water, an increasingly rare combination, while ARES needs only favorable topography. And while there isn’t much favorable topography in such places as Florida, the Netherlands and the Nullarbor Plain, it abounds in the southwestern US. Consider for example the Google Earth view below, which shows the 70km by 3o km “prospective area” surrounding the ARES Pahrump project on the west side of the Spring Mountains:

Pahrump “pediment”, showing area prospective for ARES installations

This 70 x 30km area covers an area of roughly 2,000 square kilometers where the land drops regularly at gradients of 5-10% over distances of up to 30km as we move away from the edge of the mountains, forming what we US desert geologists call  a “pediment”.  I’m now going to guess that half of this area, or 1,000 sq km, could accommodate ARES storage projects. How much ARES storage could be developed on 1,000 sq km? We can make the following rough calculations:

      • One 12.5MWh ARES system takes up 70 hectares
      • 1,000 sq km (100,000 hectares) will therefore accommodate 100,000/70 = 1,400 systems
      • Total potential storage therefore amounts to 1,400 * 12.5 MWh = 17,500 MWh, or 17.5 GWh

Now we are getting somewhere. 17.5 GWh represents approximately two Dinorwigs, and with ARES we get these two Dinorwigs from a patch of waterless desert that makes up only a tiny fraction of the total prospective ARES area in the American southwest. Potential for the development of ARES storage in the numerous surrounding “pediment” areas is clearly unlimited.

There’s just one problem – cost.

1,400 ARES systems at $55 million each will cost seventy seven billion dollars.

Okay, we can maybe take a half or maybe even a third of this number to allow for economics of scale and incremental technological improvements with time (there won’t be any major breakthroughs with technologies this mature). But we’re still looking at roughly $25 billion, which seems like a prohibitive amount to spend for enough storage to supply Nevada’s demand for only four or five days.

Although ARES might always get lucky at the tables. The city in the bottom right quadrant of the last image is Las Vegas, after all.

Site News

Readers should note that on Thursday 7th April, Energy Matters will be migrated to a new server based in the UK (previously hosted in Holland). While it is expected to be a smooth transition, there may be a 30 minute period where new content, such as comments, gets lost.

This entry was posted in Energy and tagged , , , . Bookmark the permalink.

53 Responses to Is ARES the solution to the energy storage problem?

  1. Dave Rutledge says:

    Hi Roger,

    I am rooting for ARES. From the video you linked to, it sounds like they think their entry to the market is in regulation, but the more interesting issue here is the daily or seasonal storage that is suggested by your System Ratings figure.

    I would do the cost per kWh this way. 1 metric ton of material, raised 640m would store 1.7kWh. How much would a metric ton of sand in a concrete container cost? The sand might be available from the excavation. Call it $5 per metric ton.

    We are talking $3/kWh. Tesla’s Power Wall is $400/kWh. My guess is that the sand would also last longer than a Power Wall.


    • Roger Andrews says:

      Hi Dave: From the video you linked to, it sounds like they think their entry to the market is in regulation, but the more interesting issue here is the daily or seasonal storage that is suggested by your System Ratings figure.

      I’m strongly inclined to agree with you, particularly in regard to seasonal storage. Seasonal storage involves dribbling in large quantities of stored power over long periods of time. It doesn’t require rapid ramp rates, which as you note could result in much lower costs/kWh.

      And while I don’t really know enough to give an informed opinion I can see no reason why ARES shouldn’t perform like a pumped hydro system that uses (free) dirt from rail cut excavations or concrete instead of water for weight and rails for penstocks. And as a resident of the SW USA you will know just how much land there is in California, Nevada, Utah and Arizona (and probably Oregon and Wyoming as well) that has ARES-compatible topography. And all drought-resistant too.

      I think the big problem, however, is going to be public resistance to having thousands of square miles of environmentally-sensitive desert covered in rail yards. Just think of the visual impacts. And all those endangered desert tortoises ….

  2. TheProle says:

    Increasing capacity could potentially be very cheap if you don’t want to draw it all down at once (i.e. to cover increasingly long periods of calm weather).

    if you form trains with the locomotives separate to wagons loaded with weights, then you can have storage yards at the top and bottom of the system in which weighted wagons live for months at a time if needs be, with the locomotives returning up or down the system to collect more wagons as needed.

    Wagons comprising of wheel-sets, under-frames and some blocks of concrete or similar aren’t going to be very expensive to produce – the expensive bits of this project ^^^ will be the infrastructure (track, electrical connections etc) and the locomotives (the electronics and traction packages)

    All this said, as someone who has worked in connection with heavy rail and knows a bit about what would be involved, it all seems to me to be rather optimistic and fragile – I’m pretty sure it would work, but I’m also pretty sure it will cost a LOT more to maintain in working order than it’s proponents realize (railway track takes a good deal of looking after for instance)

    • Greg Kaan says:

      Thanks for this timely comment, TheProle.

      I have seen this concept before but never with the specifications as stated in this article. When I first read it, I was going to ask if there were any railroad engineers here who could provide maintenance costs for tracks with heavily laden rolling stock.

      Would you be willing to provide some ballpark figures for this, given the stated length and loading?
      1. A single 9.2 km long track with an elevation change of approximately 640m.
      2. Six 300-ton (272,100kg) trains made up of rail cars with regenerative braking capability.
      3. 50MW peak output (I’m not sure if the braking and driving loads contribute much to wear).

      • Peter Lang says:


        What is the maximum gradient?

        What gradient can the trains manage without slipping (in worst conditions; e.g. freezing rain)

        If steeper than can be handled on regular rails, how much additional energy loss must be allowed for cog and ratchet systems or cable car systems?

        Maintenance cost for these more complicated systems?

        Total system capital cost and cost of electricity supplied from the total system of solar plus ARES?

        • nukie says:

          Depending on the relation of powered wheels to non powered wheels. With all wheels powered we have 8% here.
          With all wheels powered there would be the neccesity to have loads (Concete, dirt) and powered wheels separated to make better use otf the expensive part. The logical soulution woult then look like a container harbour, with loads (e.g. in the dimensioning of Rheinbraun, with 6 35t axis, so 210t per wagon, and maybe optimistic 180t per box) loaded and unloaded from piles with boxes onto the the power producing / consuming underwaggon.
          As far as I can see the design so far is differently. But the steeper the track is the higher is the amout of lifting and falling compared to the way just moving horizontally, loosing energy.
          If it would be designed like this, it would be reasonable to make the blocks of local stones, and locally produced concrete, most likely close to cube form. And with overhead lines not overhead but on the side, to make loading/unloading easy as Rheinbraun does it. (RWE Power today)
          Capital costs for the boxes and wear and tear of the rails+wheels are the most important points in this scheme.
          I have doubts if it will pay out, but sometimes I’m wrong.

  3. Willem Post says:


    Assuming one cycle per day, a quick estimate cost of storage is $55,000,000/(35 y x 365 d/y x 12,500 kWh) = 0.344 c/kWh.

    Ignored costs:

    – Financing and amortizing
    – Resources for O&M, and partial decommissioning, refurbishing, disposal
    – Capacity degradation due to wear and tear.

    With ignored cost: about 40+ c/kWh.

  4. willem post says:


    Correction: 34.4 c/kWh, without ignored costs; about 40+ c/kWh with ignored costs.

  5. willem post says:


    Below is calculated the LCOE of a TESLA Powerpack-based, peak-shaving system using the following assumptions:

    – The battery system is to provide 100 MWh in 2 hours.
    – The real-world loss is 20%
    – Range of charge is 79%
    – Battery degradation in year 10 is 10%
    – Replacement battery cost in year 11 and year 21 about 50% of $250/kWh = $125/kWh
    – Removal, disposal, and install new in year 11 and year 21 about 15% of new battery cost, or $37.5/kWh

    NOTE: About 100 MWh/0.80 = 125 MWh needs to be charged into the battery to recover 100 MWh, for a loss of 25 MWh/d. The annual cost of that loss is 365 x 25 x 75 = $684,375, at an assumed average wholesale price of $75/MWh over the next 30 years.

    The battery capacity would need to be 100/(0.80 x 0.79 x 0.90) = 176 MWh
    The battery capital cost would be 176 x 1000 x 250 = $44.0 million
    The capital cost of balance of plant, BOP, would be about $24.0 million
    50% is private capital requiring a return at 10%/y; 50% is borrowed at 5%/y
    The capital cost of the turnkey, battery SYSTEM would be about $68 million, or $387/kWh

    Estimates of the major annual costs are as follows:

    Year……………………………………………………………………………..1 to 10……………11 to 30
    Private amortizing batteries at 10%………………………………3,731,718…………1,865,859
    Borrowed amortizing batteries at 5%……………………………..2,845,907…………1,422,953*
    Private amortizing removal, disposal, and install new at 10%………………………279,879
    Borrowed amortizing removal, disposal, and install new at 5%…………………….213,443
    Private amortizing BOP at 10% over 30 y……………………….1,275,476……….1,275,476
    Borrowed amortizing BOP at 5% over 30 y………………………..782,297………….782,297
    Fixed + variable O&M…………………………………………………….750,000…………..750,000
    Battery system energy loss…………………………………………….684,375…………..684,375
    Miscellaneous, i.e., taxes, insurance, etc………………………….750,000…………..750,000
    LCOE, batteries…………………………………………………………18.1 c/kWh……….9.1 c/kWh
    LCOE, removal, disposal, and install new………………………………………………1.4 c/kWh
    LCOE, battery system loss…………………………………………..1.9 c/kWh………1.9 c/kWh
    LCOE, balance of plant……………………………………………….9.8 c/kWh………9.8 c/kWh
    LCOE, battery system year…………………………………………29.8 c/kWh………22.1 c/kWh

    * This cost is only for batteries; not included are the cost of removing and disposing of the old batteries, installing the new ones, and any BOP upgrades.

    Even though, battery systems can perform other services, when not in peak-shaving mode, the LCOE of the battery system, operating life of 10 to at most 15 years versus about 30 years for OCGT peaking plants, would need to become about 20 c/kWh or less to cause utilities to replace older OCGT peaking plants (which likely are already paid for) with new battery systems, unless it is mandated by law, and heavily subsidized.

    • Roger Andrews says:

      Willem: I’d be interested in your views on the use of ARES for long-term storage. Even New Hampshire has ARES-compatible topography (around Mt. Washington) and the trains would be hidden in the trees.

      • willem post says:


        I think ARES costs/kWh will be too high, as I showed, with little prospect for technical breakthroughs.

        Recent developments in REDUX flow batteries in Germany have utility-scale potential at very low cost, because of the use of low-cost materials.

        The Whiteface Mountains, NH, has Mt Washington with a shoo-shoo train to the top and a road to drive to the top. Inaccessible during the winter.

        Mt Washington is known for the fiercest weather in the Northeast, top wind speed 238 MPH.

        New England would be much better off by building HVDC lines to Canada, which has at least 5000 MW of extra hydro capacity, due to projected demand that did not happen, and is building at least 5000 MW more.

  6. Peter Lang says:

    Just to provide some context for the claimed 1 GWh plant:

    Figure 10 here: shows GB would need about 8 TWh of energy storage (i.e. 8,000 x 1 GWh plants) for weather dependent renewables to paper the GB electricity system.

    Here is a conceptual study (not viable) I did some time ago for an 8 GW, 400 GWh pumped hydro plant connecting two existing large reservoirs in the Australian Snowy Mountains Scheme:


    1. Where could GB site 8,000 1GW ARES plants (or even 1,000).

    2. What’s the total land area required and what’s the land acquisition cost and ongoing rates?

    3. What’s the total land area of solar and wind farms plus ARES that would be needed to power GB

    4. What’s the total area of nuclear plants needed to power GB?

    • robertok06 says:

      “4. What’s the total area of nuclear plants needed to power GB?”

      This is easy: scaling from French data, even with a low average capacity factor of only 75%, in order to generate 300 TWh (the remainder in hydro + other) one would need 3/4 of 63 GW, i.e. ~48 GW, or 12 power stations with 4x 1 GWe reactors. Each of the 12 power stations covering 1 km2, for a total of 12 km2.

      The same land area covered with PV panels would generate in the UK 12E+6*5*8760=0.53 TWh of intermittent electricity, i.e. 300/0.53=570 times less than nuclear.
      Wind? At 2x 2MW turbines/km2, the 2x2x12=48 MW of wind, even at 0.35 capacity factor would generate 48E+6*0.35*8760=0.147 TWh, i.e. 300/0.147~ 2000 times less.
      With bigger turbines, 5 MW at 1/km2, wind would generate 20% more, i.e. still 1650 times less than nuclear.

      Highest power density source, waiting for anti-matter annihilation or DT fusion to come online… can’t beat it.

      • Peter Lang says:


        Thank you for this.

        For other readers, please ignore this bit in the part of the comment quoted by Robertok 06 “(i.e. 8,000 x 1 GWh plants)”. It’s nonsense.

  7. Nador says:

    “But ARES has a major advantage over pumped hydro too. Pumped hydro needs favorable topography and water, an increasingly rare combination, while ARES needs only favorable topography.”
    I do not think this is true. One can use closed reservoirs to avoid evaporation. And I think it is cheeper to build two big tanks for water and a pipeline, than to build a lot of railways. This whole ARES thing seems extremely expensive.

  8. Leroy Essek says:

    Here are a few technologies that could be used to increase overall efficiency in rail transportation and shipping. First is a company called Joi Scientific that can use any type of water as a self sustaining hydrogen on demand technology. Instead of a fuel cell combust the hydrogen and oxygen gas at 5,800 F inside a zero emission steam boiler called the Hydrogen Technologies Inc company located in Stockton California. This boiler is so clean it qualifies for carbon trading credits. The steam from combusting hydrogen and oxygen gas is super critical 580 psi pressure that could be used to generate electricity at only two cents per KWH operating cost. This electricity could be used to power the electric motor of the train propulsion. Another great engine the American Hydrogen Associations facebook page posted a article about was the Koenigsegg Freevalve Camless Engine. This engine can run on multiple fuels including steam. Not only the Freevalve Engine could be used for road transportation but it could also be used to power a electrical generator. Many are highly doubtful about the astounding claims of Joi Scientific located at NASA’s Kennedy Space Center. Even though GoPro Co-Founder invested $5.5 million into Joi this currently held proprietary technology now has the money to obtain international patents in 40 different countries. Once the patents are secured Traver Kennedy the CEO of Joi Scientific will show and tell how this transformational technology can help many people throughout the world. One such place is the Salton Sea in California. Look up Agess Inc Facebook page and read the letter of support by Joi Scientific to help desalinate ocean water and generate the lowest cost energy in the world.

  9. Syndroma says:

    12.5MWh is approximately 1.4 m3 of petrol. 1GWh is approximately 3 tank trucks of petrol.

  10. Peter Lang says:

    3. 50MW peak output, 12.5MWh storage with the option of scaling up to 1 GW.

    4. Capital cost $55 million ($1,100/kW installed. I can find no information on costs/kWh.)

    Here’s a simple back of an envelope estimate of LCOE for the entire system (generation, transmission to storage and storage):

    $55 million for 12.5 MWh energy storage capacity = $4.4/Wh = $4.4B/GWh
    [Check with Roger’s figure: $77 billion for 17.5 GWh = $4.4B/GWh]

    To get 1 GWy of energy per year supplied from solar PV plus ARES, we’d need sufficient solar PV capacity to supply say 1.25 GWh of energy every hour of the year (assumes 25% energy loss). In winter that would require sufficient solar generating capacity to provide all the power almost every day and sufficient storage to last say 10 days with little generation (say 250 GWh energy storage). On simple assumptions we’d need approximately:

    10 GW of solar PV @ $2/W = $20 billion
    10 GW transmission capacity with average line length say 200 km @ $1,000/ = $2 billion
    1 GW ARES with say 250 GWh energy storage capacity @ $4.4B/GWh = $1,100 billion
    TOTAL capital cost for 1 GW generating capacity, 1GWy energy storage capacity system = $1,122 billion = $1,122/W!

    That is, ~200 times the cost of nuclear to provide equivalent firm dispatchable capacity! [hard to believe – is this correct?]

    And no destruction of thousands of square kilometres of little desert critters.

    [Please check these back-of-an-envelope calculations.]

    • Peter Lang says:

      Woops, I’d intended to do the LCOE calculation but then headed off in a different direction to calculate the capital cost of the full system to generate 1 GWy per year.

      I’ll calculate the LCOE later after the discussion on this estimate has run its course.

      • willem post says:


        See my above comments and calculations per kWh just for building the storage system. I estimate at least about 40 c/kWh, which is prohibitive.

        There is no need for complication by combining the 40 c/kWh with the location and weather-dependent generating cost, c/kWh, of a PV or wind energy system.

        ARES is a complete pipe dream.

        “This polymer-based redox-flow battery is ideally suited as energy storage for large wind farms and photovoltaic power stations,”

        Read more at:

        • Peter Lang says:

          I agree its a pipe dream – my figures, if they are correct, show why.

          I disagree you can calculate the LCOE of the storage alone. I think you need to calculate the full system cost if the power source is to be unreliable, weather-dependent renewables. The capacity factor depends entirely on the distribution of the power for charging – e.g. in winter and in long periods of overcast weather or sand storms. And the capital cost of the system depends on the energy storage capacity and the generating capacity of the renewables has to be matched with the energy storage capacity to provide the power demand under worst case weather conditions. I think calculating LCOE of the storage system without taking into account these factors would give a meaningless result.

          It would be entirely different calculation if the energy supply was from a reliable source. However, in that case you need to include the cost of the ‘fuel’ for storage – i.e. the cost of electricity purchased for storage.

          • willem post says:


            Of course, I agree with you, as any sane, experienced, energy systems engineer would.

            As a preliminary calculation, taking the turnkey system capital cost and dividing by the lifetime energy, kWh, obtained from storage, provides an immediate insight regarding feasibility.

          • Peter Lang says:

            I agree your calculation gives a minimum cost of energy for the storage component of the system. But the real cost will be much higher if powered by weather-dependent renewables. The plant has to buy sufficient electricity each day to fully recharge the storage for full drawdown that night, and do this every day-night of the year. Therefore, it has to buy power at peak time and sell at off-peak time. The purchase cost of energy is huge And the sell price low. Alternatively, if you don’t use a full cycle every day and, instead, store for sustained periods of low insolation, then the plant will have a very low capacity factor thus the cost of electricity will be huge.

            Any way you look at it, this is another loony idea from RE advocates who haven’t even bothered to do the most basic reality checks.

        • Greg Kaan says:

          Thanks for the link, Willem. It looks like they are serious about commercializing the battery.

          The thing that rarely gets mentioned with these new storage concepts is that they would also apply to traditional base load (CCGT/coal/nuclear) baseload thermal generators to allow them to run at higher capacity (by storing power at low demand periods to be used as a peaking source). I saw in a television program that Dinorwig is used in conjunction with the AGR reactors in this manner.

          Similarly to what Peter stated, you would need to factor the operating savings of the generating plant into the cost of the storage system to see if it was worthwhile

          • Peter Lang says:

            99% of all electricity storage globally is pumped hydro. No other technology is close to being viable. Pumped hydro was viable when powered by cheap, off-peak power from baseload plants in the 1970’s and 1980s. However, new pumped hydro plants are rarely viable now. All the other energy storage ideas that keep popping up like monsoon frogs are even less viable. When the costs per kW and per kWh energy storage capacity are not provided, its usually a sign that they are vary costly.

            And storage is much more economic with reliable baseload power that is cheap in off peak times and the stored power can be sold at peak times. With cheap reliable power you can use the near the full storage capacity every day. Six hours for storing (e.g. midnight to 6 am), gives around 4.5 hours generation at full power. This gives about 20% capacity factor.

  11. Gaznotprom says:

    Thanks again for the great analysis.

    My conclusion (for what it’s worth) is the best and easiest storage energy system is the one that involves Hydrocarbons.

    These gimmicks, are corn thrown to the green-set to discuss and virtue-signal to others and pontificate ‘we have the renewable solution…’

    When we and many know, thanks once more to all the analyses on this site know – yep these work, yep solar works, wind mills work, tesla works… But the question is (lost on many) is are they viable, they are not free, and can we scale them and can our economy function effectively…

    Really is: ‘it’s the economy stupid!!!’

  12. Alex says:

    My first thought is that it would be easier to dig reservoirs and have pumped storage. Yes, water is only about one third the density of rocks, but storage volumes are easier. If you have to conserve water, then it’s easy to enclose the water.

    Something like this might be needed:
    (though I think that system is too ambitious, needs too much HVDC, and would risk salinating a chunk of Scotland).

    The downside of Ares is also in the cost – $1,100/KW. Batteries are now at about 600, and could come down to 200, so the lifecycle cost advantages might go. In most of Europe the cost would be higher still (no empty mountains – can’t disturb the skiers).

    One energy storage technology that could undercut for very large scale storage all is Isentropic

    Take a 200mx200m area of dry rock land for your hot store, drill tubing down to 200m, and you have 8 million cubic metres for heat. Do the same for the cold, and you have about 500 – 1000 GW hours of storage, with the heat pumping station in the middle.

    Easier though to just build nuclear power. The UK only needs 100GW or so of capacity. About 20GW would shut down over the summer and storage is mainly catered for by the thermal mass of electrically heated buildings.

  13. Olav says:

    Electric storage is very expansive unless naturally filled reservoirs which we have in Norway are available.
    Heat storage is less costly and it gives some interesting possibilities as 1 m3 water heated 1 C equals lifting the same amount 400m. Here you are heating up to 400C but concrete has less heat capacity than water but it is heavier so it cancels out.

    If you have excess heat at low demand times like some power plants (all nuclear and many fossil fueled plants) has then there is a possibility of storing some of the production.
    Heat to Heat storage -> electricity has very low losses and is most useful if storage has high temperature (400C) and it is used every day.
    For industry that is using heat is the efficiency even better. If you are among the few that use electricity to produce heat for your process. Very few do so now unless “cheap” hydropower is available. Excess heat can be produced and stored at low price times and retrieved from storage when electricity price is high. Other industries which closes down in evening can store heat during the night making next morning startup less costly. This has been done at some extent in industry already but the water storage used is limited to 100C. With solid storage it is possible to store heat at 400C and possible above and then a new “game” is on… Almost 60% of industrial heat demand is satisfied with heat at lower temperature than 400C. We should not concentrate all on moving electricity production away from fossil fuels. Making part of heat demand from an alternative source has the same value and it is less costly.
    As seen below is the heat storage footprint very low, allowing a location within power plants perimeter which is close to power source and allowing dual use of turbines, generators, grid and skilled operators.

    Rated / maximum capacity 1.25 GWhth / 1.65 GWhth
    Volume / mass of storage medium (Heatcrete®) 30 800 m3 67 700 ton
    Footprint 4 400 m
    Height 14 m
    Procurement and construction cost estimate USD 50 million

    Cost is off cause an issue as always: 1,25GWhth is only 0,5 GWh e. A 20th of the PHS capasity below without “long” storage and at 1/8th cost. So this storage is more costly than PHS but the location and grid issues are solved. If the stored heat is used s heat then cost compares with PHS.

    One PHS in UK as sourced from Euan.
    Machine hall volume 211,140 m^3
    Tunnels 16 kms
    Reservoir volume 6.7 million m^3
    Head 542-494 m
    Power 1.8 GW
    Energy stored 9.1GWh
    Duration 5 hours
    Cost £425 million

  14. I’ve been looking into how many trains an ARES system would need to store enough electricity to smooth out seasonal solar variations to the point where a meaningful amount of year-round solar “baseload” generation could be sent to the grid.

    Using 1TWh of storage as an arbitrary threshold and the storage = mass x gravity x head x efficiency formula I find that 350,000 trains, each 200m long, each weighing 2,000 tons and moving up and down 650m hills would be needed to do this.

    These numbers seem unreasonably large. Maybe I’ve slipped a few zeros. Could someone please check them out for me?

    • Dave Rutledge says:

      Hi Roger,

      I would do a power rather than an energy calculation here. Assume that an ARES system for the US would need a capacity of 300GW. The average US grid demand is 500GW, but I assume that it would not be still and cloudy everywhere in the country, and that you could overbuild on wind and solar.

      For locomotives take the power of a locomotive to be 3MW. This would be 30,000 locomotives. For comparison, the US fleet is 25,000 locomotives.

      Take the descent rate of the trains to be 1m/s. The required mass would be 31Mt. At 100t/car, this would be 310,000 cars. For comparison, the US fleet is 1.4 million cars.

      Reference for fleets, US Office of Industries


    • Stuart Brown says:

      I make moving that mass that distance to be 1.24TWh assuming no losses of any kind. Roughly a shipping container is 100m3 on the outside and 15m long. That’s close to 3000 containers full of concrete at 2400kg/m3. Or at 10 containers per train to give 150m of container to fit into your 200m train, that’s 300 trains. I think a cog might have slipped with one of us!

      • Stuart Brown says:

        Drat. Do not post while under the influence. 3 million containers, 300,000 trains. You’re right Roger and Dave

    • robertok06 says:

      “Using 1TWh of storage as an arbitrary threshold and the storage = mass x gravity x head x efficiency formula I find that 350,000 trains, each 200m long, each weighing 2,000 tons and moving up and down 650m hills would be needed to do this.”

      1 TWh is 3.6E+15 Joules;
      2000 tons is 2 million kg, times 10 (g) times 650 m equals 13 billion J, times 0.72 for roundtrip efficiency (like pumped hydro, 0.8 pumping times 0.9 generating) gives 9.4 billion J.

      Dividing, 3.6E+15/9.4E+9=383 thousand trains… so, I’d say you were almost right, Roger. 🙂

      That’s a scary number, isn’t it?… and it doesn’t bode well for this new technology.

  15. Jonathan Madden says:


    I agree with your numbers, with 26% efficiency loss. It comes as no surprise of course when one remembers that a pint of hot tea requires electrical energy sufficient to raise an 80kg individual to the top of the Post Office Tower in London, 200m!

    I am just wondering what happens to a nuclear power station when a trip occurs and load drops to zero in a matter if seconds. Is there a built-in mechanism that can absorb power from the generating turbines while the core is moderated, which I presume takes at least a number of minutes if not longer?

    • Stuart Brown says:

      The (fission) power output of a PWR drops off very quickly when the control rods drop in – to 10% ish within a couple of seconds. That last bit comes from the decay, which then drops off more slowly. It’s generally enough to keep the circulation going to cool things down and dump the heat out of the cooling towers/river/sea water rather than into the turbines – as is normal for 60% of the actual energy produced anyway.

      Back off to check my maths!

  16. Gentlemen: Thank you for all those responses. I’m glad I seem to have done my sums right but dismayed by the massive amount of rolling stock that would be needed to keep what is actually a small-scale ARES seasonal storage facility running. (In a recent post I estimated that 7TWh of seasonal storage would be needed just to replace the 3.6GW of nuclear generation from Hinkley Point C with “baseload” solar). The conclusion seems to be that ARES needs far too many trains to store a useful amount of energy for a long period of time, although it could of course still be used for short-term load following and grid stabilization.

  17. Peter Lang says:

    To get 1 GW constant power we’d need:

    Say 20 GW solar PV @ $2/W = $40 billion
    20 GW transmission (average length say 200 km) @ $1/ = $4 billion
    Say 250 GWh storage capacity (for sustained period of low power) @$4.4 billion/GWh = $1,100 billion
    Total cost = $1,144 billion = $1,144/W
    c.f. Nuclear = ~$6/W
    Total system cost for solar PV + transmission to storage + ARES is about 200 times more than for nuclear (to supply a constant 1 GW power)

  18. Peter Lang says:

    Here’s another energy storage proposal:

    As the “Energy system of the month” SolarServer presents an extraordinary suggestion by German physicist Prof. Dr. Eduard Heindl, Professor for e-Business Technology at Furtwangen University. Heindl designs a hydraulic hydro storage as a cost-effective alternative to conventional pumped storage power systems, and identifies potentials and risks.

    A new option could be hydraulic hydro storage, presented in this article. The basic idea is to lift a large rock mass and to store the potential energy. If necessary, the rock mass is lowered again and potential energy is transformed into electricity. Approaches that try to move large masses by mechanical means, such as ropes or tracks, failed due to the cost per stored kWh. However, there is the possibility to lift a large mass by hydraulic means, and therefore this path is an interesting choice. A cylinder of rock, preferably granite, is separated from the surrounding stone, within its natural environment. This is done by wire saws as they are used in quarries to separate large stone blocks. In this case, the wire saw is designed to cut a cylinder wall and the cylinder bottom off.

    500m high by 1 km diameter = 1614GWh!

    This is the current daily electricity production of Germany.

    Questions (especially for those with any experience with engineering in rock masses):
    1. how do cut a perfect cylinder 1 km diameter (with sufficiently wide cut to allow for strain caused by the release of horizontal stresses)
    2. How do you hold the cylinder of fractured rock together as one perfect cylinder and how do you hold the walls in place.
    3. How do you maintain the seals between the rock cylinder and the walls as the cylinder slides up and down
    4. How will the raised cylinder react when Rayley waves pass through the area (one for geophysists 🙂

  19. Greg Kaan says:

    This thread is turning into complete nonsense, not due to the commentators here but simply through the “solutions” being presented to try and cope with intermittent power production.

    On another site (Peter Lang will know which one), I was casually dismissed for stating that mass grid level storage would have been developed and deployed already if it were possible (for baseload peak shaving as mentioned earlier). That worthy felt sufficient efforts had not been made previously due to lack of need and that practical solutions were only a matter of time (and presumably money).

    I still stand by that statement – we need a fundamental breakthrough in physics to solve intermittent power production while maintaining the ability to meet demand as required.

    • I’d go further. I’d say we need a divinely-inspired breakthrough to solve the intermittency problem. A modern-day deus ex machina

    • Peter Lang says:

      Greg Kaan,

      Yes,”I know what you are referring to and the individual involved – he also used to troll here but seems to have given up, thankfully.

      I agree with your comment that “we need a fundamental breakthrough in physics to solve intermittent power production while maintaining the ability to meet demand as required.”

      I’ add – don’t wait up! It;s taken 215 years for electricity storage to get to the state its at. Learning rat is very slow. If there was a likely prospect on the horizon it would be used first in conventionally powered submarines. Most are still using lead-acid batteries.

      The battery powered submarines run for hours between recharges. Nuclear powered submarines run for over 30 years on a single fuel load. So, why on earth are we mucking around advocating anything else – and delaying progress?

  20. Olav says:

    Smoothing out seasonal solar variations is almost impossible, Only big naturally filled resevoirs can do so and even that works ony when you have a small population combined with high elevation land where ice age carving has made the reserviors for you,
    PHS is only viable if dayly use is possible..
    Wlllem made the following easy calculation for Ares “Assuming one cycle per day, a quick estimate cost of storage is $55,000,000/(35 y x 365 d/y x 12,500 kWh) = 0.344 c/kWh.”
    If we do the same for a typhical PHS then we get 0,00364 c/kWh. Ares/PHS gives then 94,55.
    Ares is almost 100x worse than PHS which is struggling with economics today. Tesla Powerwall (assuming 7000 dollar cost when installment,taxes and inverter ar added in) gives 0,3427 c/kWh which is same bad economics as heawy rail cars on a slope .
    This is why 99% of man made storage is PHS and other storage types will struggle to reach 10% up from 1 % now, I see , however, a possibility in the concrete storage I posted a link to abowe.
    That storage gave 0,00548 which is 1.5x PHS

    • Peter Lang says:

      Smoothing out seasonal solar variations is almost impossible, Only big naturally filled resevoirs can do so and even that works ony when you have a small population combined with high elevation land where ice age carving has made the reserviors for you,
      PHS is only viable if dayly use is possible..

      Indeed. And that needs to be repeated, repeatedly, ad infinitum!!

      However there is very limited sites available for future pumped hydro, so it will provide a decreasing proportion of global electricity storage.

      However, there’s no need to worry. The world has effectively unlimited energy stored in nuclear which can remain easily stored at very high energy density until needed – just like the fuel in your car except that nuclear fuel is 20,000 times more energy dense when used in a conventional reactor or up to 2 million times denser when used in breeder reactors. Energy security and energy storage are no problem for the future – we just have to get over the public fear and political blocks to progress.

    • Peter Lang says:

      Dave Rutledge said:

      I would do a power rather than an energy calculation here.

      I disagree. If the problem we want to address is to make weather dependent renewables dispatchable and baseload capable, then the issue is primarily energy storage capacity, not generating capacity.

      Olav said:

      Smoothing out seasonal solar variations is almost impossible

      Olav is absolutely correct. This illustrates the magnitude of the problem:
      Solar Power Realities Supply-Demand Characteristics, Storage and Capital Costs
      Web post and comments here:

      This is a simple, limit analysis of the solar PV capacity and energy storage capacity that would be required to supply the Australian National Electricity Market with power to meet the 2010 demand profile. It is a limit analysis in that the profile of power supplied is scaled up from a single commercial PV power station in Queanbeyan, NSW (at latitude 35 S, 200 km inland). The power readings are at ½ intervals and continuous for 2 years.

      Figure 7 (in the PDF linked above) shows the capacity factor for continuous periods of 1, 3, 5, 10, 20, 30, 60 and 90 days. The lowest capacity factors were in winter and were:

      5 days = 4.3%
      10 days = 5.7%
      20 days = 6.6%
      30 days = 7.8%
      60 days = 8.6%
      90 days = 9.4%

      Average energy consumption in winter is 600 GWh per day (average 25 GW). At the capacity factors listed above the solar generating capacity (GW) required to provide 25 GW average power per day for different amounts of energy storage would be:

      5 days = 686
      10 days = 524
      20 days = 448
      30 days = 383
      60 days = 347
      90 days = 315

      The least cost option is with 30 days pumped hydro storage (if sites were available, which they are not), i.e. $2,800 billion – versus $100 billion for nuclear to supply the same average power (Figure 10).

      The land area required would be 11,000 km2 (8,000 km2 for pumped hydro reservoirs and 3,000 km2 for solar PV) versus 26 km2 for nuclear plants. The CO2 emissions would be 20 times higher than with nuclear.

      • Dave Rutledge says:

        Hi Peter,

        I did the $/kWh of storage calculation in the first comment of this post.

        Obviously the pricing has to be done both by the power capacity (W) and by energy (J).


        • Peter Lang says:


          I did see your first comment but I thought you were joking. I posted a reply, but it was deleted. You didn’t actually provide a cost figure ($/kWh storage capacity). Could you add your estimate? It should be capital cost divided by energy storage capacity.

          The issue is storage, not power, so the primary issue for comparison with other options is the cost per unit energy storage capacity.

          However, I suggest, for the cost comparison to be useful we need to compare the cost of two systems that meet demand and other requirements of the electricity system. For example to be dispatchable and baseload capable we need to include the capital cost of the generators, transmissions from generators to storage, and storage.

          Using Roger’s numbers, I very roughly estimated the cost of storage capacity for solar PV generating capacity plus transmission to storage plus 250 GWh of ARES energy storage at $1,144 billion = 4.5/Wh = $4,500/kWh.

          This seems wrong (an order of magnitude too high). What am I doing wrong?

  21. Bernard Durand says:

    I propose to use plutonium instead of concrete, because it is 7 times denser, and then, you can store this dangerous matter in deserts.
    And to store jumps in electricity produced by marine wind turbines in case of gales, I suggest to install such systems on cliffs

  22. Michael says:

    It’s shocking that such a project even exists. What a bizarre concept, lining the sides of hills with thousands of tracks and running these things up and down them. It seems to me like such a waste of human ingenuity.

Comments are closed.