Compressed Air Energy Storage - How viable is it?

Energy Storage - Compressed Air

One of the most critical aspects of the implementation of renewable electricity is the ability to store electricity.  If a good solution existed right now, our situation would be a good deal easier.  On the face of it, compressed air seems a likely candidate: relatively easy to make, store and use - so what is the problem?  Why isn't it used routinely?

More Thermodynamics than You Ever Wanted to Know?

We usually speak of storing and using energy without being very precise about what we mean.  That ends forever if you take a few chemistry or engineering courses.  Thermodynamics rules everything.

Let's start with the usual definition of work - using a force to push something a given distance (in the direction of the force).  The amount of work is the force multiplied by the distance, and has units of energy.  If we lift a 1 kg mass by 1 metre in the earth's gravitational field on the surface of the earth, then the work done on it is the force required: 1kg x 9.8 m/s2 (9.8 Newtons), times 1 metre, or 9.8 Joules.  Since a Watt is 1 Joule per second, then in principle (no friction), this lift could be carried out in 9.8 seconds by a 1 Watt electric motor.  At the end of the process, the weight has acquired 9.8 Joules of gravitational potential energy.

We just constructed an energy storage device.  The weight we lifted could now be allowed to descend, giving its potential energy back to an electrical generator and making electricity in the process.  This is in fact the basis of possibly the most effective existing way of storing electricity.  Water is pumped from a low reservoir to a high one at times when there is a surplus of electricity, and then allowed to flow back when there is a shortage.  For useful amounts of energy storage using reservoirs that are not too large, one generally requires reservoir height differences of a hundred metres or more, which limits this to suitable terrain.

So what about compressed air?  Surely a cylinder of compressed air contains energy that could be used to drive something?

This is where it all becomes a little strange.  The energy content of compressed gas isn't very different from that of uncompressed gas at the same temperature.  For an ideal gas, the energy contents are identical.  How come we can get work from the compressed gas?

The answer is that compressed gas has a lower entropy than the uncompressed gas, and that the amount of useful work you can get out of something when it changes depends both on the change in energy content and the change in entropy.  We usually focus so much on the energy side of things that we ignore the entropy side.

If the compressed gas has no more energy than the uncompressed gas, where did the energy used to compress it go?  The answer can be found in the old bicycle pump experiment.  When you compress a gas it becomes hot.  In fact all the work put into an ideal gas to compress it is turned into heat.  If that heat is thrown away, the same amount of energy as was in that work is thrown away with it.

To look at a definite example, if we take 1 cubic metre of air at 1 atmosphere pressure and 20C and compress it to 10 atmospheres pressure, its temperature will increase very considerably - to 293C.  If we want to store this compressed air at 10 atmospheres pressure and 20C, then more compression will be needed as we cool the gas, or its pressure will drop as its temperature does.  The total work done on the gas, and the total heat lost are both about 91.7 Watt-hours (Wh).  (This assumes that the air is an ideal diatomic gas.)

This gas would now have a lower entropy than the same amount of uncompressed air.   The entropy change is 796 J/K (Joules per degree Kelvin).  Note the units are energy per degree.  This gives a hint of how the entropy change is related to the work that can in principle be extracted from the compressed air.  That work can be calculated by multiplying the entropy change by the temperature of the environment in degrees Kelvin.  20C is 293K, so the amount of work that can in principle be extracted is 233 kJ, or 64.8 Wh.  If we compare this with the work done compressing the gas, we see that the efficiency of the process is about 71%, even if the compressor is perfectly efficient.

Looking at the expansion of the same air back to 1 atmosphere, using a motor to do work in the process, we can work out that the temperature will fall to -121C, and that the work that is done would be 47.5Wh.  The efficiency of ths process is thus 47.5/64.8 = 73%, even with a perfect motor.  The round-trip efficiency for energy storage and use would then be just 52%.  With real compressors and motors it would clearly be considerably worse.  These numbers above are for a compression ratio of 10.  If we instead use a compression ratio of 100, things get worse still, with a round-trip efficiency of 27%.

This actually gives a clue as to how to improve the situation.  The maximum efficiency of the cycle depends on the pressure ratio, and rises to 100% as that ratio approaches 1.  The answer is to use staged compression, with cooling back to ambient temperature between the stages, and staged expansion, with reheat back to ambient temperature between stages.  If we get the 100 times compression by two stages of times 10 each, then half the work goes into the first stage and half into the second, with efficiencies as for 10 times compression - a huge improvement.  If we use four stages (ratio 3.17), then the maximum effficiency would be 72%.  If we take into account that real compressors and engines are not perfect, and neither are coolers and reheaters, we can see that real overall efficiencies achieved are never likely to be very good, even with very complicated equipment.

Whether technology is useful depends, though, on comparison with the alternatives.  The overall efficicency of a compessor train and a compressed air car may not look all that high, but an internal combustion vehicle engine can look pretty inefficient, even with North American fuel prices.  This means that an air-powered car may make some sense.  For more details on the MDI air car, see some MDI engine tests.  Notice that in a conventional car you get free heating, but in a compressed air car you get free cooling.  

Bulk power storage is another matter.  Large reservoirs of compressed air can be and have been constructed, but they are not used simply to drive engines to regenerate power.  Building large heat exchangers to warm the air in a power generating unit would be very costly and not very efficient, so the air is instead heated to a much higher temperature before the expansion turbine by burning natural gas in it.  The whole installation is thus a sort of gas turbine, with the difference that the compressor and power turbines are run at different times instead of together.  This is no longer a straightforward energy storage device.

Libelle -

Nicely done treatment of the subject.

It should be clear by now that compressed air storage is analogous to a very inefficient spring, with considerable losses in both compressing and releasing the spring.

Heating the compressed air prior to expanding it through a turbine is just a way of putting back some of the energy that was lost during compressive heating. As such, most compressed air torpedoes of the WW II era employed such heating to increase their speed and improve their range.

As I see it, the only way to make large-scale compressed air storage even halfway viable would be if one had a ready use for the wasted heat of compression (such as for space heating or process heating perhaps). Or possibly if one could store the some of the heat of compression and then give some of it back during the expansion part of the cycle. Either way, we're talking about high capital cost in relation to what would be gained.

Compressed air storage hardly looks like a winner.

many small towns in the us and canada have water towers. some of these towns are dying a slow death.

so it would seem that they might have excess capacity of stored water.

just wondering if any could be used to store potential energy, at night say, and used to generate electricity during the day ?

Pumped water has been used for years to store energy. As the article states, it's not about efficiency, but about alternatives: What are the options, what are the characteristics and costs of these options?

A water tower is not very big when talking about municipal power generation.

but the potential energy can be used:

Pint-size hydro power on tap. Rentricity, a start-up in New York City, has come up with a hydroelectric generator that lets municipal water facilities generate power.

Granted, they are using excess pressure within the system to run the generators. Nonetheless, I have seen city water delivered at 80 PSI (had to install a pressure reducer)!

We can rapidly calculate the gravitational energy stored in the water using the following equation:

Energy Stored = mgh

m = mass of water
g = acceleration due to gravity, 10 is a reasonable approximation
h = height water is stored above your energy extraction device, e.g. turbine.

A typical power station has a rated output of 1000 MW, if we assume it runs at an average of half capacity (500 MW) for 24 hours, this is a total energy output in this time of 43.2 GJ. Lets say we have a 200m high tower. Using the above equation, we would need to pump 21600000000 kg (21.6 million metric tonnes)of water into the tower to store the same amount of energy. For reference, 1 kg of water is 1 litre, 1 american gallon is 3.78 litres, making it 5714 million american gallons. This is impractical, even if we have many towers.

P.S. if anyone notices a mistake in this quick calculation feel free to correct me!

500 MW running for one second is simply 500 M joules.

Running for 24 hours it is:

500 M x 24 x 60 x 60 = 43,200,000 M joules = 43.2 T joules

If the tower is 200 meters high then you need to store this much water in one day:

43.2 T / (200 * 9.8) = 22 G kg

Since one cubic meter contains 1000 kg water, therefore it means:

22 G/1000 = 22 M cubic meter water

I assume that you need to store in water half the energy you need and the other half is used right away. Then you need 11 million cubic meter water.

If you have a dam as big as an approx cricket/football/hockey stadium (which I take as a square with each side one stadia or 200 meters) then you need a depth of these many meters:

22,000,000/(200 * 200) = 550 meters

I don't know the construction cost of that big a stadium. In Pakistan the cheapest brick is in villages, its size is 6" x 4" x 3" and come at a price of atleast Rs. 5 per piece. That is when its made on stoves that are heated by burning tires which is a very polluting way but we get along with that in village where air is very clean. We also have the advantage of having low prices. The labour of the brick factories hardly earn enough to feed themselves.

Since one brick is 24 square inches at maximum area and stadium is:

(200m x 550m x 4) + (200m x 200 m) = 440,000 + 40,000 = 480,000 square meter

Since there are about 40 inches in one meter therefore it means an area of:

480,000 * 40 * 40 = 768,000,000 cubic inches = 32,000,000 bricks = Rs. 160 million

This is when the dam's boundaries that is its walls and floor is just 3" thick. That obviously is not enough to hold a pressure of 550 m high water column at floor and 200 m long horizontal water pressure at walls. I assume we need atleast a ten ft thick wall. 10 ft = 120" = 40 walls in a row at walls and floor. The cost now is:

Rs. 160 million x 40 = Rs. 6.4 billion

I assume it to be Rs. 12.8 billion assuming 50% rise in expense each for labour cost and mortar cost.

Pakistanis on average consumes 220 watt energy and has a per capita gdp of $880. Gdp density is about 0.25 watts per dollar or 8 million joules per dollar. Since the cost of dam is Rs. 12.8 billion or 200 million dollars so it means an energy expense of:

200 M x 8 M = 1600 M M joules

The dam is supposed to provide storage for 250 mega watts for a duration of 30 years, the usual life time of dams. In that time the dam would store:

250 M x 30 x 365 x 86400 = 237,000 M M joules

The amount of money spent to make the storage place is just this much percent of the gdp gained in the process:

1600/237,000,000 x 100 = 0.675% (that is less than one percent)

The base thickness of a dam wall is proportional to height of the wall. For a stable gravity wall you would be looking at the width of the base being a third of the height. This gives you a factor of safety of 1. I wouldn't want to live under a dam which can only just hold back the pressure of the water. FYI the three Gorges dam wall is 101m high and 115 m wide at the base. Essentially a factor of safety between 3 and 4.

FYI - you can't make a dam wall out of bricks. Bricks have no tensile strength which is why brick houses fall down during earthquakes. A brick wall of a dam would burst without being reinforced with steel tie rods.

The most efficient structure to store water is a circular tank. The wall of the tank can be extremely thin because the tank wall works only in tension ie hoop stress. Tanks walls can be steel work or steel reinforced concrete - any material that has a high tensile capacity.

Water as an energy storage device has been around since the first civilisations. It is a common misconception that it needs to be a great differential between the input and output height to actually produce power. A 1m height difference with a very large volume will give the same energy output as a 100 m height difference with 1/100th of the volume.

The Dutch and other low lying countries have worked with this for years. Frankly the world could do a hell of a lot better than looking into the past for energy creation and storage solutions. A low lying paddock, a paddock a little higher, a few windmills, a few inclined screws and and a water source you have yourself a very powerful energy generator and storage device. Even better if the paddocks lie beside the ocean as you can utilise the potential of tidal flows to fill it up when the wind isn't blowing.

Or possibly if one could store the some of the heat of compression and then give some of it back during the expansion part of the cycle. Either way, we're talking about high capital cost in relation to what would be gained.

The capital costs of CAES are relatively low, and heat storage isn't that expensive either. Cleverly designed AACAES might not cost more than pumped storage and could be almost as efficient.


That is an excellent essay; concise and very informative. I wish you had written my physics textbooks!

I get the feeling that Libelle and I were thinking along the same lines, but she obviously wasn't reading me or this article would have been different.

BTW, one term which wasn't mentioned here but probably should have been is "availability".  This is the amount of energy which can be recovered from a system, and it's one of the crucial concepts I took away from my thermo course (which changed the way I saw the world as it did Libelle).

A suggestion. Why don't you or PG or someone do a thermo primer? Libelle did decently, but as you said to analyze this you have to use the Helmholtz free energy and/or exergy concepts.

a thermo primer

I always reckoned that almost everyone would switch off when the equations started.

to analyze this you have to use the Helmholtz free energy and/or exergy concepts.

Exactly. To do this in a way that is accurate is not trivial. I've made a couple of presentations to economists, with a positive reception, but I don't know how much stuck.

but she obviously wasn't reading me

It's "he", btw.

Sorry.  I can't tell the players without a program, even when I'm one of them.

Libelle is a very feminine name, most likely caused by the affix 'elle'. It's the German word for Dragonfly isn't it?

It's the German word for Dragonfly isn't it?

Yes. I used to fly one of these:

Amazing little machines.

One of the original "glass slippers".  Part of the Aptera's parentage, from the looks of it.

Well, now, all this is good as far as it goes, but it dropped the real ball early on- water storage. Pumped hydro storage. Been done for centuries. Works just great. DOES NOT NEED A HILL. Look it up.

Please, please, people, I ( and others, too) have said this over and over and it never seems to stick. YOU DON'T NEED A HILL TO STORE PUMPED UP WATER. All you need is an ELEVATION DIFFERENCE between two pools (or, if you want, lakes) of water. Whether the elevation difference is between a lake up a mountain and ground level, or a lake on the ground and another one 200 meters down a hole makes no difference. The energy stored is the same either way- and it's just as accessible and just as efficient.

There, I've said it again.

OK, so you will say holes are expensive. Well, yes and no. Some are, some are already there ready to get full of water. And people dig big holes all the time and are smart about how to do it.

I envision a sustainable energy system in which every source is a water pump. Windmill water pumps, solar thermal (my personal passion) water pumps, everything pumping water up from down or up from here, storing gobs of energy. This gives us an artificial big powerful river of water. We know how to use big rivers of water, right? A big water turbine (very efficient, very well known) driving a big alternator (very efficient, very well known) 24 hrs/day, all the time.

On my morning walk today I was playing around with the idea of replacing that expensive gear box in those 2megawatt windmills with a water pump. Much cheaper, much more reliable, and lighter by far than the gearbox and alternator that is now up on that stick.

And don't tell me it would freeze. Absolutely no way. Too big, too much heating even tho very efficient.

Your suggestion sounds good in theory, Wimbi, but is there a working example of a pumped water storage scheme using a big hole (as opposed to a mountain reservoir) in the real world? Don't get me wrong, I'm not opposed to the suggestion, but would like to see it being tried and working successfully someplace in the world. It's hard to imagine that it would be entirely overlooked if it was practical. There are flat places in the world that have lots of wind but no reservoirs for pumped storage (ie the Texas panhandle).

I can see some significant problems in trying to implement pumped storage using a hole in the ground:

1) you'll need an awfully big hole, actually, an underground reservoir (cave system?). Otherwise, the hole will fill up with water very quickly, maybe in a matter of minutes.

2) the underground reservoir may fill up with ground water, not just the water you're pouring into it. You'll have to constantly pump out ground water, using energy in the process and negating the whole idea of using this as a pumped storage area.

3) you need an above ground reservoir to store the water that you're pumping up

Just punch in "pumped hydro below ground" to your favorite search engine and you will get the same string of examples I did. Take your pick.

BTW. reason pumped hydro is better is that unlike gas-air- it is incompressible, or nearly so, and the energy you put into pumping it has to go to elevation change, and not partly into bumping molecules around faster at random and then having them go bump something else outside and waste that energy (heat of compression, heat transfer to willynilly).

You can hardly put any energy into water by just merely pushing on it while it is sitting there, any more than you can put energy into a concrete wall by pushing it. It doesn't move. No move, no energy. But you can put a hell of a lot of energy into air by just pushing, because it moves (compresses), as any tire pump with a blocked exit will tell you. And you don't get it all back, not by a long shot.

That's what the post said, of course, more elegantly. Me, I am a rude, crude engineer, totally
inelegant. Just make it work, any which way you can.

AACAES should be able to get 70-80 percent round trip. Good enough I'd reckon. Pumped hydro can be a bit better with modern tech, perhaps 90 percent round trip.

Underground pumped hydro is really an interesting underdog technology. Possibly the cheapest, one of the most efficient, and environmentally friendly storage methods.

90 percent? Single point, yeah, but the reversible pump/turbines have rangebility problems. At least the ones at Power Vista in NF used to.

I'm thinking low head. The Dutch have a plan that envisions damming the inland lake:

Modular pumped hydro would scale the size of each module for effective operation. And, of course, with multiple modules, there is less need for a wide range of operation between minimum and maximum flow for each unit.

Ah, the light bulb comes on...

When is a giant hole in the ground a liability?
Answer: when it is an exhausted open pit mine.

When is a giant hole in the ground an asset?
When you can use it to store/extract kinetic energy.

Thank you wimbi

When is a giant hole in the ground an asset?
When you can use it to store/extract kinetic energy.

Actually you need two reservoirs - one substantially above the other - in the real world, not many exist like that.

The problems of storage of energy in any gas/fluid are simmilar to those causing peak oil.
What is required is an adequate and cost competitive flow rate of energy (both charging and discharging), which will be determined by the need for suppliers to make a profit and the abilty of consumers to afford the cost.

This was a good simplification of a difficult subject Libelle, most people don't understand entropy and it's implications. I would like to see somebody fill up an MDI vehicle and drive it continuously at normal highway speeds (not just a demo around the block) and see just how far it gets - not as far as most people might think is my guess.

Just for info, here is a pump storage system in Wales with the delightful name of "Ffestiniog Power Station"

Yes, that is a very good real world example of what is required - and why pumped water isn't adequate as a non-FF battery for the UK at least.

360MW (from a store of 2 million cubic metres of water at a rate of 27 cubic metres per second) for a few hours (about 20 by the looks of it) is a small part of 820 MW UK declared net wind capacity in 2006, supposed to be ~25GW by 2020.

It is said that this unit could supply the needs of the whole of North Wales for a few hours - not surprising since it is one of the most mountainous and least populated parts of the UK - sadly nowhere near what is required if we want smooth non FF power for the whole of the UK.

now that you mentioned it, i dont know why an abandoned oil well wouldn't work for this application. many wells are capable of producing thousands of barrels (of water per day) and they will also accept thousands of barrels per day under injection under just the hyrdorostatic pressure of the water column.

it would seem that the problem arises in designing a pump and electrical system to move that much fluid from a significant depth. one would imagine that to be workable, the pump would have to operate variably, according to the electrical energy available. that problem could probably be worked out, but at what efficiency ?

and an added bonus would be the heat that could be harvested from such a system.

it would seem that the problem arises in designing a pump and electrical system to move that much fluid from a significant depth. one would imagine that to be workable, the pump would have to operate variably, according to the electrical energy available. that problem could probably be worked out, but at what efficiency ?

Moving fluid from large depths is best done by placing the pump on the bottom of the resevoir rather than on the surface. This probably makes the water pump-in-windmill idea a bit less attractive.

Variable operation costs a bit in efficiency, but it's quite acceptable.

actually, i was thinking submersible pump and submersible generator. not in the same well though.

Ah, OK. Wimby brought up the pump in windmill idea. I don't think that would be very practical anyways.

NPSH is a bitch.

One thing to think about, through, is that the turbine is at the bottom of the connection between the two reservoirs. In the case of a mountain or foothill, this is convenient ... because generally the bottom of the mountain is more accessible than the top, and the fact that the top is just a passive reservoir and a pipe is how you'd rather have it anyway.

If the top reservoir is at ground level and the bottom at the bottom of a hole 200m deep, then that location at the bottom is an inconvenience.

Mind you, it does not mean its technically infeasible, it just increases the cost.

By the same token, a top reservoir on a tower is certainly feasible ... but putting it on the ground where the ground is at an elevation saves the capital cost of the tower.

And the 100m is not a technical feasibility requirement ... its also an issue of capital cost. If the reservoirs and pipe and turbines are all sized to provide a good balance between volume per unit and scale economies ... then the power storage of that standard unit increases in proportion to the elevation between the top and bottom reservoir, so twice the elevation(NB) is twice the energy storage per cubic meter of water capacity.

That is why the 100m threshold is referring to a purpose-built facility ... if there is an existing hydro generating facility, and the capital cost of the reservoirs are already covered, and the generators can be converted to reversible pump/generator for a reasonable capital cost, then that is a potential target. However, modular pumped storage avoids the problems of interfering with the river ecosystem.

(NB. Strictly speaking that is a slope twice as steep, to get twice the elevation with the same capital cost for pipe.)

What the other poster said, NPSH is a bitch...

Pumped hydro storage. Been done for centuries. Works just great. DOES NOT NEED A HILL.

But you need a surface body of water at one end or the other (or your cost goes way up), and that's a problem.

Take the Ludington (MI) pumped storage facility as an example.  The reservoir is quite large; it's plainly visible on Google Maps.  This reservoir is not suitable for fish habitat or recreation because it is filled and emptied rapidly.  Yet as large as it is, it can only supply a fraction of the state's electric needs for a matter of hours.

Trying to scale up facilities like Ludington would mean lots of real estate devoted to such single-use reservoirs.  In many parts of the country, it would mean large lakes where there are now none, and large evaporation losses where water is already scarce.  This does not look like it can or should be our model.

CAES has its issues (mostly related to efficiency), but scalability and siting are not among them; the atmosphere is far more vast than our ability to pump air, and it is not subject to local depletion or scarcity.  If the heat of compression can be partially stored and recovered, the efficiency can be improved.  Biofuels or biofuel byproducts (e.g. tail gas from "green diesel" F-T plants) can supply the energy for reheating air.  The same underground caverns you suggest for pumped hydro will store far more energy as compressed air, and the surface footprint of a CAES plant is minuscule.

I was once a CAES skeptic.  I ran the numbers, and now I'm an advocate.


What are we to make of the document at "MDI engine tests"? I have trouble understanding it. I wonder if you know what they mean by 'autonomy' of a vehicle. What is the 'global concept' that permits significant autonomy?

I remember reading somewhere that compressed air energy storage was being used, or maybe being planned to be used, where the pressure vessel was a closed natural cave. The pressure rise was quite small and consequently the efficiency was much higher. But, of course, caves can't be carted about in a one or two passenger vehicle.

I think that by "autonomy", they just mean how far it will go, and by "global concept" they mean the design of the car.

"20C is 297K, so the amount of work that can in principle be extracted is 233 kJ, or 64.8 Wh. "

I couldn't get the numbers to calculate, and then I remembered that 0C=273K, which would make 20C=293K.

293 *.796 = 233

So should it read

"20C is 293K, so the amount of work that can in principle be extracted is 233 kJ, or 64.8 Wh. " ?

Oops. Thanks for the correction.

Should probably give apples-to-apples comparisons:

Pumped hydro - energy in/energy out ~80%

Batteries - ~70-80%

Compressed gas - ~60-70%

Investment costs - compressors and expanders cost more than pumps and turbines, but less than battery or capacitor banks. OTOH, a reservoir costs a whole lot more than a hole in the ground.

Modern pumped hydro with state of the art equipment can be ~90% efficient round trip.

Some battery technologies are >95% efficient round trip. But lose a bit over time. Altair's product looks good though.

Caps are >99% efficient round trip (due to lightning charge/discharge, they would melt if they weren't extremely efficient). Also lose a bit over time. Eestor has to hurry up...

Diabatic CAES is actually the lowest capital cost storage system currently available. However, that's not entirely fair, as CAES 'cheats' the definition of storage by using significant amounts of natural gas. They do use a whole lot less than CCGT especially SCGT so make sense in eg grids with large amounts (%) of wind but also in nuclear baseload grids.

The capital costs of AACAES are higher but look affordable. 70-80% efficiency looks possible. Problem is no one's built a large commercial system yet, so we have to wait and see how that goes.

I don't think holes in the ground are cheaper than resevoirs. Costs I've seen are 0.1-1 USD per kWh(e) capacity. Can you build a pumped hydro hole in the ground for less than that?

Per the Ridge Study, the fuel-to-electric efficiency of even non-regenerative CAES can be on the order of 80%.  This makes biofuels extremely competitive (80% is as good as direct-carbon fuel cells but requires no new technology and can operate on any combustible fuel or fuel byproduct).

I've been wondering how the economics of high efficiency bio-energy diabatic CAES compare with the adiabatic thermal storage variant.

On the one hand, the bio-energy variant has lower capital costs than the heat storage system. On the other hand, the heat storage system doesn't use chemical fuel so avoids fuel costs altogether and frees chemical feedstock up for other uses.

Costs for a large AACAES version appear to be somewhere around 1 USD per Watt, roughly twice the cost of a bio-energy variant. But still quite low.

The bio-energy CAES does have lower marginal storage costs than AACAES. This suggests maybe an optimum situation where the AACAES would be used as regularly used diurnal load follower (maybe 5-10 hours), while the biofuel CAES is used for slightly deeper backup (perhaps tens or even hundreds of hours). This also mitigates the issue with sensible heat loss from the AACAES thermal store. (another way to do this is thermochemical storage eg dissociated ammonia reaction, but this looks expensive with current tech).

I think that to get this sort of efficiency from compressed gas, you would have to keep the gas hot. It could be done, but I don't know of any actual examples. Do you know of any?

Here's a quickie I did using COCO simulator. Five stage compressor, each stage with adiabatic efficiency ~85%, four stage expander, isentropic efficiency ~85%, all inlet temps at 70 deg F. Power in/power out = .50

Use more expander stages and/or heat the expander inlets and you get over .6+

The Espcinc website has some concepts that get very low heat rates (4000 Kj/kWh or almost 90% isentropic efficiency) and also an adiabatic system as well:

The above is pretty much with off-the-shelf machinery. Using regenerators or BAHXs instead of shell-and-tube HXs together with custom designed wheels (90% adiabatic efficiencies) you can do a lot better.

What did you assume for pressure loss and capital cost of the heat exchangers?  Those are big factors.

Intercoolers on a typical compressor (centrifugal and positive displacement) would be shell and tube, with the air in the shell. Pressure drops would be in the 1-2 psi range, 1 psi being on the higher pressure units. The coolers would usually be provided by the compressor vendor, and could be 5-10% of the cost.

An alternate would be to use trayed or packed direct contact coolers. Low pressure drop and small approaches. But you'd probably not be able to buy a compressor that way, you'd have to pipe it up yourself.

When I was looking at stuff after writing Useful questions Re: CAES, I found references to rock as a heat-storage medium going back decades.

The earliest air separation plants, vintage 40-50s, used gravel filled regenerators for heat transfer.

I can see why that didn't stick. Too much parasitics, too large a storage volume. Magnesia firebrick gravel might work. Structured beds are best. Corrosion is an issue though.

So clearly this article is leaving out Advanced Adiabatic CAES.

A little short sighted don't you think?

To improve the efficiency, RWE and GE are working on a new design called advanced adiabatic CAES (AA-CAES), in which the heat that is removed from the air during compression is stored and later used to reheat the gas as it is discharged.

"In this case, the air is hot enough to drive an air turbine without using combustion gases," Marquardt said.

The efficiency could be increased to 70 percent, and if combined with wind power, an AA-CAES system would release no carbon dioxide.

RWE and GE are currently doing a feasibility study looking in particular at what material would be best for storing the immense heat. Marquardt thinks the likely choice will be ceramic bricks, but a possible alternative solution is a bed of rock pebbles.


And while we're on the subject of Viability.

How about we start mentioning things like Capital Cost comparisons.


And while we're at it, how about we notice that making huge swaths of artificial lakes for pumped hydro, takes huge swaths of cement. OH NO CEMENT PUTS OUT CARBON! THEREFORE IT'S USELESS!! (equivalent hyperbole)


All in all, this is a rather silly article, where the claim that efficiency is everything. When in reality, thats not true.

Good points, was thinking the same. I wouldn't suggest rock pebbles. They appear attractive because of low cost, but the thermodynamic properties are unfavorable, and experience with solar thermal heat storage systems in the US has shown that this typically increases the levelized cost of the heat storage system. More parasitic losses, larger storage volume, more insulation costs... not good. For high temperatures, magnesia fire bricks are one of the best options. Steel also works quite well for higher temps, but is expensive. Perhaps if something like a large amount of used ball bearings (pebble bed heat storage) can be outsourced for a low price, this will be very interesting.

For low temperatures, I'd suggest water. Cheap, environmentally friendly, and excellent thermo properties. Low temps are promising because they put less demands on the equipment. Engineer Poet has some good thoughts on this, hopefully he has some time to comment.

Classic regenerators used firebrick. Alumina or ceramic packing was also used if fast response was desired (large surface area).

Steel is OK, but some sort of packing would be better (Raschig rings rather than balls).

Some kind of structured bed packing would be best from a parasitic losses viewpoint. However, I think lower temperature operation (and storage) offers distinct advantages over high temperatures. The use of more conventional turbomachinery and H2O thermal storage could make a very cheap, reliable, and environmentally friendly system.

This isn't very convincing. They've run mass-transit streetcars in France and many mining engines on compressed air machines.

They used to run compressed-air LOCOMOTIVES(capable of hauling +500 tons rail loads on level) (instead of steam) for coal mines, etc. They operated at much lower pressures(~1000psi) than our current pressure vessel technology is capable of(~3000 psi). In terms of volumetric energy storage air compressed to 3000 psi inside a tank can hold as much energy (.1 MJ/L) as most batteries which are also heavy and complex to make. The efficiency of the old 'compound' compressed air locomotives with atmospheric interheaters was around 25% which is really not so bad, especially compared to the electric cars which are getting so much buzz.

According to the U. of Minnesota, it takes 8 electric hp at the compressor to get 1 compressed air hp at the end use:

And this is at fairly low pressure - probably less than 200 psig in an industrial plant. Running your car on 3,000 psi cylinders will be less efficient, as outlined above.

Of course the next question and/or stage of development that would suggest itself is, If your going through the trouble of compressing air, why not just go ahead and liquify it and use cryogenic storage?

This is all the more true if you are already using expensive multi stage heating and cooling of the air.

This is not a new suggestion, and all you have to do is key in the words "Liquid nitrogen economy" on Google and you will see that mountains of research has already been done. Technically it's not a problem, but whether it is viable or not on a cost basis remains to be seen, especially as flow batteries such as sodium sulfur and the Vanadium Redox batteries continue to improve for bulk stationary storage (utility scale).

But as oil and gas prices rise some of the alternative storage methods, including cryogenic air or the liquid nitrogen economy may begin to look viable. The development of energy storage is absolutely crucial to the development and growth of solar, wind and wave energy. Any of the above (battery, compressed air or cryogenic air) seems to hold much greater promise than the once greatly touted "hydrogen economy".

The main point is this: Storing energy is not thermodynamically free and it will never be so "don't let the perfect be the enemy of the good". But it still must be done if we are to break out of our current energy paradigm. It truly is a comparative/competitive game and may the best method win.


Right now, LN2 is cheaper than gasoline in $/gal. More expensive if you go by energy content.

Ugh, Thermo. Entropy was always JUST outside of my reach of understanding. This is a very good point though of the heat generated (and lost) when air compressed. Or if one uses a 'can of air' to clean the computer keyboard, it quickly gets very cold. I have even seen frost on CO2 cartridges used in BB guns. Temperature shift is energy and that energy must come from some where.

Perhaps this heat can be used productively. As the storage cylinder is filled, it also pre-heats a tank of water for another use. Or the discharging cylinder in a vehicle is used to provide cooling for the passengers (a plus here in Florida). Of course, when you fill up your tank, you get all that heat back. The biggest lesson in thermodynamics: There is no free lunch!

Yes, but the whole point of AA-CAES is that when you compress the air, it gets hot.
So if you could hold onto that heat long enough to when you decompress it, then you wouldn't need an additional heat source.
(As was previously mentioned)

As for the energy, the energy is primarily coming from the sun. Which is essentially a fusion reactor. Thats not "Free" energy, but it's certainly plenty for the next couple billion years.

Of course. Energy and mass are inextricably tied to each other via the brilliant E=MC2. So no energy is 'free'.

I've been thinking that some geothermal resevoirs could do quite well in providing the heat source, even lower grade heat (say 400 K) would be suitable. Solar thermal might work too if there's good direct insolation resource available. Just mirrors and thermal storage. This will allow traditional proven CAES systems to be used.

How about using the waste heat from compressing air to run a stering engine pump to store water in a reservior. You won´t capture all the heat but you may get enough to make electricity to heat the compressed air on use.

The whole point of the multistage heating is to generate as little heat as possible, which will lead to maximum efficiency. A sterling engine requires a minimum temp difference (and is not that efficient).

Besides you could simply us the heat of the compressed air to heat the air before compression, which would do about the same with probably higher efficiency.



*blush* fixed. completely missed that.

Just for those of you who would like a better understanding of the engine, MDI has a joint venture with IndraNet Technologies of New Zealand. The company is called
IT-MDI -Energy. They have a great video showing the MDI Compressed Air Engine with Multi Fuel External Burner Demonstration from January 2008.
Direct from the test bench at the MDI factory in Nice, France. Learn about the MDI compressed air engine with external fuel burner.

Really high-level info there. Main advantage is "no internal combustion"?

I've done the math before, and the keynote article is correct. The big killer of CAES is the thermal losses. Eg. if compressed gasses were of much use for power (eg. MDI's "compressed air engine"), then high-pressure hydrogen fuel storage in cars could extract some significant fraction of their required power from an expander driven by the fuel coming out of the tank. Not so. Do the math yourself. (can collect perhaps 1% to 2% of the chemical energy content of the fuel, cretainly not worth the added weight.)

I suspect "compressed air powered autos" are a scam and/or entirely dependent of fuel to re-heat the compressed air. May as well use a closed Rankine cycle engine (Steamers). Waiting to see any company disprove that.

"So what about compressed air? Surely a cylinder of compressed air contains energy that could be used to drive something?
This is where it all becomes a little strange. The energy content of compressed gas isn't very different from that of uncompressed gas at the same temperature."

Can't run this with uncompressed gas...Bird Mk 7, 50 PSI pneumatic powered ventilator

A cool 2200 PSI compressed gas powered rocket...something I've always wanted to do...
Extra damage cuz it is powered by oxygen...starts at 4:20

The most efficient turbo diesel is what, around 25% efficient? And gas engines significantly worse, at about 15% in real-world applications. In either case, most of the 75-85% wasted fuel goes bye-bye as truly useless heat.

How about using that waste as the re-energizing heat source when decanting a bottle of chilled compressed air?

With an auxiliary compressed air motor, vehicles could use much smaller and fuel miserly conventional engines running at peak power (for greater efficiency) and call upon the compressed air motor for extra foot/pounds when needed. Deceleration force could even be used to recharge the compressed-air tanks. Because air is being compressed periodically in real time, insulation would reduce heat loss in the air tanks, at least in the short term (i.e. during the current period of engine operation.)

A little complex, but perhaps no worse than a Toyota Prius, and anything that improves vehicular fuel consumption has gotta be a REALLY good thing. What would EACH 1% of efficiency improvement be worth on a daily basis, measured in barrels of crude that would NOT go up in smoke?

The best five-story high marine diesels have an efficiency above 50 %.

Hi Starvid,

I wonder if the efficiency of the very large marine diesel is due to the displacement increasing at a cubic (exponent of 3) rate while the surface area of the combustion chamber walls (primary vector of heat loss) increases at only a square (exponent of 2) rate? Just an idea (I'm in no way an engineer), but if that is so then the smaller engines are doomed to greater ratios of heat loss, which in turn might offset, somewhat, the advantages of dinky engines.

You mean friction, right? (direct radiative heat losses aren't that big, especially not in diesels with lower average engine operating temps).

Area to volume is one thing. But more important is the compression ratio. The most obvious way to improve a diesels' efficiency is to increase the compression ratio. But there's limits to this due to how much materials can take. Increasing the compression ratio requires a bigger, heavier engine.

In a marine application, weight is less important because of it's relatively small impact on fuel use, so the increase in engine efficiency easily outweighs that. And there's often plenty of space too. That's why 50% peak diesel efficiency is possible today, and perhaps 60+ percent peak in the future. And long distance shipping is done near the peak efficiency of the engine (actually, near the peak efficiency of the entire ship, which can be a bit different, technical but also economic).

Not so in automotive applications. Increasing the weight has too much impact on fuel economy, and size constraints are more stringent. Plus there's greater variance in RPM so automotive engines have to have good average efficiency too. The hope is for more exotic, stronger materials to be developed, but IIRC that's advancing only slowly due to higher cost, durability concerns and complexity. Drivetrains have seen some impressive advancements though. Continuous variable transmission, 7 speed automatic etc. And serial (plugin) hybrids can operate the generator at peak efficiency so we'd expect those to be very efficient considering the efficiency of electrical generators and advanced batteries/caps

Do you have info on the average HP that the largest container ships run at, as a % of rated capacity? I read that 80% of rating is typically the peak efficiency, which suggests that a 80MW rated engine would normally operate at 64MW. On the other hand, as fuel becomes more expensive, one would expect lower speeds, as consumption is the square of speed.

What are you seeing these days?

As a rule of thumb, fuel use goes up to the 3rd power of velocity, so substantial savings can be had by reducing the cruising speed a little. However, there are technical issues with going too slow. For example, the ship geometry and engines/drivetrains are built with an optimum range in mind. Another variable is of course how much cargo is on the ship, and how much that cargo is worth (specific value). If the cargo is worth a lot of money, odds are the captain will go full steam! (interest charges and stuff) So the answer is probably: it depends...

Existing ships do appear to decrease cruising speeds much more often the last few years. There was an article in the Guardian a month ago that mentioned trains and aircraft were also slowing down:

Looks like they're about 20 ish percent slower than before. I'm not in the shipping business so don't have accurate up to date figures on MWs, but I'm guessing 80% of max power is probably too high for most shipping practice with today's bunker oil prices. Many ships would have issues with cruising speeds below 40-50% of max speeds. But that's just speed not power. Can't find any good power/speed curves for container ships...

Thanks. I'm thinking about estimating fuel consumption - perhaps an estimate of 50% of max engine power rating would be reasonable.

My goal is to evaluate the % of fuel consumption that could come from PV - so far it looks like it could be significant. It certainly would be cost-effective.

Any data on normal fuel consumption for large container ships? I've looked, and so far haven't found much.

50% would definately be reasonable for low value products.

But when you're shipping 2000 TEUs full of Ipods... you'll want to deliver those quick. Plus there's JIT requirements sometimes; if the premium paid for JIT is larger than the marginal fuel costs, or if the shipper is contractually bound to JIT, then reducing engine power poses logistic limitations.

Wind is a no-brainer for ships. Can't say that it's unproven! With modern kites, it looks like ships could typically get 10-30% of power from wind. Maybe even more in the future.

There's new ship designs that look very promising. The axe cleaver type for example.

I'm waiting for direct carbon fuel cells to finally arive commercially. Ships could get most of their power from wind and solar, with the rest provided by efficient DCFC. Coal could be used, in the future detorrified biomass might be used.

As for data on normal fuel consumption, yes it's difficult to come by. This smaller ship uses 40 tons worth of fuel per day @ 20 knots. 1.6 - 1.7 tons per hour, but it's a pretty small ship with 168 meters x 27 meters, leaves less room for PV. Can't be more than a MW peak of PV. If the ship travels at (say) 10 MW continuously to save some fuel, then 10% of peak capacity PV, maybe. 2-3% PV powered? If high efficiency PV is finally coupled to low cost... the holy grail... then maybe the situation will improve. Think nano-antenna converters.

It looks like 90% of max power is standard by design, odds are with today's oil prices they won't be doing that, so 10 MW looks more reasonable to me. If they can get even less then things look better for PV.

There's a thread here with discussions on this type of issues:

300-350 metric tonnes of fuel oil per 24 hours is mentioned for a big ship (Emma Mærsk)

I wonder about very light hydrofoil kind's of outriggers (horizontal extension), or sails (vertical), to carry PV panels. I should think a ship wouldn't have to be limited by the sq surface on it's deck.

My experience on ships is that sunshine comes from above, and from the waves - that would be something to take advantage.

Just some random thoughts.

Need some breakthroughs on ultra efficient ultra cheap highly flexible PV for sails.

Outriggers? Durability of the structure is a serious issue. Wind speeds on the high seas can be unforgiving. No good if they can't be deployed when the wind blows hard.

PV on the side of the hull to capture wave reflected photons?

Maybe one of those kite things pulling the ship could be covered with PV as well.

Making the ship twice as long doesn't nearly result in twice the fuel use, but double the area for PV. This is not very practical though. What kind of out outrigging do you have in mind?

Tubercles can make rudders more efficient.

It's all risky business if you ask me. But ships don't really have a choice with peak oil now do they?

"Need some breakthroughs on ultra efficient ultra cheap highly flexible PV for sails."

Well, I think cheap & flexible are here, though highly efficient it's not. Might be able to do something with it, though.

"Outriggers?.....No good if they can't be deployed when the wind blows hard."

Well, the need to pull it in during high winds would reduce utilization - the tradeoff between structural strength and utilization would require design work and analysis.

"PV on the side of the hull "

Absolutely. Put it just about anywhere you can, I think.

"kite things pulling the ship could be covered with PV "

Sure. Large surface area, and thinfilm PV is pretty light.

"What kind of out outrigging do you have in mind?"

I'm not sure - something hydrofoil/catamaranish, I'm thinking, either on spars to the side or towed.

I'm sure basic PV on the superstructure, deck and hull would work - the real question is raising the surface area for a greater contribution.

If you've got cheap and flexible PV why don't you tow it behind the ship?

I think that would make sense.

Cheap and flexible PV (thinfilm) is less efficient than crystalline silicon, so you'd have to do some design work to evaluate the tradeoffs between flexibility and greater surface area needs.

PV has been cost-effective for ships for only a short time, when the curves of rising oil cost and falling PV cost met. A lot of design work needs to be done, and I suspect it will progress slowly until ship designers are confident that this cost structure is permanent. In fact, I haven't heard anything about it from the industry yet, though they have to be thinking about it.

One other thing that might be useful for big ships is this:

Net energy density greater than bunker oil!

That looks promising for any long-distance transportation need: air, ship, or truck.

That sounds great! Just a huge roll on the stern, rolled out when on the open waters.

Sure - just roll it out: 50M wide, and 1500M long, and get 15MWp.

Yes, and that may be doubled or even tripled with future nano tech...

There would have to be a cleaning system (or person) to get rid of the salt that remains on the PV after evaporation though. And of course, dragging the roll behind the ship is going to require a bit of extra power.

D.Benton_Smith -

The idea appears to have some merit.

It all gets down to how easily (and cost-effectively) one can store and then move waste heat from the compression phase to then subsequently be recovered during the expansion phase. It appears to be a not insignificant heat transfer and storage problem. But probably not insurmountable, though it would appear much easier to achieve in a stationary power plant than in a prime mover for a vehicle.

However, my gut feel is that advanced electrical storage systems will eventually prevail over any compressed air system, regardless of how clever it is in moving waste heat around. Some concepts eventually prove to be technological deadends, and I suspect that compressed air energy storage might be one of them.

I know of some air compressors for industrial gas production that have been running for 40-50 years, with regular maintenance. Heat exchangers also.

How long do batteries last, again?

Shameless plug for a favourite "ugly duckling" technology:

Nickel-Iron batteries (Edison Cells, Ni-Fe) seem to have an "indefinite" life.

Wikipedia gives a life of "more than 20 years"; there are stories of Edison Cell batteries in fork-hoists running for over 40 years with no loss in battery capacity. Compare that to Lithium based batteries at around 4 years, maybe.

Ni-Fe batteries withstand massive overcharging and complete discharge, so charge-discharge control circuits can be simple (=cheap). Their raw materials are abundantly available, (relatively) low cost and of (relatively) low toxicity, and the batteries operate at room temperature with virtually no explosion risk (=low risk of burns, thermal or chemical).

The disadvantages are that charge-discharge efficiency is low compared to other battery technologies, at around 65% vs 90-95% for Lithium; per-cell charge and discharge currents are low; and they are as heavy as lead-acid batteries. I wouldn't power a car with them in their current form.

But for stationary applications, their robustness, durability, long-run cheapness and safety must make them worthy of consideration. Now that someone has dramatically lowered the cost of photovoltaic cells by inventing a way of printing them, the efficiency penalty of Ni-Fe batteries is smaller than it was.

(I see the Wikipedia article has been updated to say they are being evaluated for wind and solar power systems.)

One of the links in the Wikipedia article says 90% efficient when new, and about 80% efficient when older.

Nickel costs more than fifty thousand dollars per metric ton, and increasing the demand substantially will increase the prices even more. That's a serious restraint.

And 100 W/kg power density really sucks for automotive apps. That's a metric ton of batteries for a 100 kW vehicle.

Maybe nano technology can improve nickel iron batteries? Only the power density has to be increased for use in an efficient serial (non-plugin) hybrid. But it's way to expensive for bulk stationary energy storage.

Hi Joule,

The longer I study your reply the more I am compelled to simply agree with it completely.

As liquid fuels decline the issues of gas storage keep cropping up in different settings, particularly in relation to vehicles. First of course was the 'hydrogen highway' until cycle inefficiencies became clear. The next likely setting may be compressed natural gas as a liquid fuel alternative as advocated by Pickens and others. While both hydrogen and CNG are meant to be combusted unlike compressed air the common issues are range and weight. For example I don't envisage micro-cars with heavy CNG tanks.

All this talk about how efficient a particular system or energy source is,should be placed in perspective to
whats being used now. Electrical generation and delivery weather by NG or Coal or Nuke has enormous waste. Take a fluorescent 48 inch lite bulb and hold it
in thin air, below low hanging high tension wires and
watch it glow white hot. The power leaking from those
high voltage high tension lines has never provoked such
scrutiny...yet everyone is holding all other alternatives to higher standards.
Building Nuke or Coal or NG electric plants are expensive as is transportation of oil around the worlds
seas..ditto for coal or liquid natural gas.
The infrastructure to transform into energy and distribute this finale product is expensive in every form we use now.
I quess that being forced to do something as opposed to
wanting to do it, plays into the psycology of it.
I myself would cringe in horror if I were forced to
spend my time with a young and beautiful women.Knowing
that I would be wasting calories better spent on such
things as napping and nose picking.

To this point, no one has bothered to mention or recognize air tools. The discussion shunted into powering cars and transportation, but consider all the jobs that air tools do now. Air tools are mature, well-developed, and already in widespread use. Compressed air is a different way to match "impedance" than electricity. I could see a compressor on a windmill or on a 12 volt PV array pumping up a tank for small scale use. In many respects I'd find that much better than a bank of batteries.

Scale matters.

cfm in Gray, ME

Further, in off the grid residential applications, one of the bottleneck energy users is refrigeration ... an application where compressed air offers a more direct approach than electricity->heat exchanger.

One can imagine a solar PV system with batteries for electronic devices and high efficiency LCD lights, and a wind / compressed air system for things that can be driven directly by compressed air.

Air tools are not particularly efficient.  They gain in high specific power, low cost and elimination of shock hazards, which is a worthwhile tradeoff for such small devices.  The same tradeoffs are not worthwhile for e.g. highway vehicles, which have the burden of storing their energy as well as converting it to motion.

Refrigeration can be performed by expanding compressed air, true.  However, the efficiency is low (and Hilsch tubes are far worse; the only reason they are used is when efficiency needs to be traded off against light weight, cheapness and reliability).  Storage of compressed air is bulky compared to batteries, has low efficiency and presents an explosion hazard if not maintained properly.  While it may be possible to use compressed air, it would make more sense to use immediate surpluses of energy from wind and the like to make ice and hold that for refrigeration (CoP > 2) instead of building two separate energy systems, one of them very lossy.

Mass displacement storage of surplus energy for fixed applications seems to be the most promising. Pumped water has been much explored, both here today and elsewhere.
The above Wiki article is a good primer for us non-engineers as is Libelle’s. However, Libelle’s article mentions in its intro actually lifting a weight to store energy moving on to relate this to pumped water storage. We have a very common geared mechanism the uses a lifted dry weight to store energy, the coo coo-clock mechanism. Energy from a human arm is used to raise a weight and some of this is recovered to drive the clock and the bird, dancing milk maids or whatever you have. I have been toying with investigating the usefulness of a similar mechanism to store the surplus output from PV and wind installations. Unfortunately I haven’t had the time to learn the physics and engineering involved to draw any conclusions. It seems to me that it could be a relatively low tech, compact and inexpensive way of storing electrical energy. With deference to the hole in the ground water pumper above (thanks wimbi) what would be needed for a typical installation would be an electric motor/generator, a low speed gear mechanism connected to a spool/cable/weight suspended over a not too deep pit. The whole thing is variable in scale and could have a power output from a few watts to many kilowatts. Even individuals with relatively small PV wind installations could use it. What needs to be investigated is what mass and gear ratios would be the best and if in the end enough energy could be stored to make it worthwhile. Even though there would be the usual maintenance issues these should be less than for a fly-wheel mechanism. The advantage over pumped water is in the eliminating of the pump and turbine and the need for less space. If anyone has some more detailed analysis of this, pros and cons please pass it on.
There is some superficial discussion of the idea here (link below) although a bit different that what I have been thinking :

Once you run the numbers you quickly realize that gravitational storage for domestic purposes doesn't work.

Suppose you have a 2 KW Solar PV which on average generates power for 6 hours a day. That's a very nice sunny spot! OK so you have 12 KWHr's (a bit less than my typical usage) of which you need to store say, 6 KWHr for night time use.

6 KWHr's is 6x3.6x10^6 joules or around 20 MegaJoules of energy.

Suppose you have 10 meters of vertical drop (which is a very big tower or hole in the ground). Gravity gives you a factor of 10, then you need a weight of 20x10^6/(10*10) = 200,000 kg's or 200 tonnes!

So this is why nobody is selling gravitional storage systems for domestic PV.

Thanks TT. I assume you were just implying 100% efficiency too. For your 6 KWH a 5+ meter cube of concrete or a 3 meter cube of iron would be needed. Your average SUV would only be good for about 60 wh of energy storage but maybe we could crush a hundred of them into a handy cube. The same reality applies to pumped hydro so we are talking about a large amount of water to be moved for practical application. Using the turbines for pumps in the Niagara gorge or other high dams might be practical using excess grid power but it seems micro-storage wouldn't be that practical either. Seems my grand kids are heading for a life of only doing certain things like they used to be done, when the wind blows and the water flows.

Using pumped hydro for micro-storage (that is, household scale) is not on the cards in any event, both because there are substantial scale economies involved, and power stored is proportional to elevation ... a 200m elevation stores four times as much power per tonne of water as a 50m elevation. Add the scale economies together with the fact that few people live on the sides of mountains, and its not a household-scale dispersed storage system.

Once a suitable elevation is located, then the alternative to a single enormous facility is a large number of smaller facilities, produced in large enough volume to begin to benefit from scale economies in production, while each is large enough to benefit from the scale economies that apply to operation.

However, its easy to over-estimate the amount of storage needed to accompany a given amount of wind turbine generating capacity. As we scale up from local to regional to national networks, the availability problem from the harvesting of volatile sustainable renewable resources can be reduced in the first instance by connecting regional grids together with HVDC trunks ... since the power availability of a wind farm is greater than the power availability of an individual turbine in the farm, the power availability of several wind farms in a region is greater than the availability of a single wind farm, and the power availability of multiple wind farms in multiple wind resource regions is greater than the power availability in any single region.

And the back-up for the medium term is, of course, the existing power generation system. It is not, after all, power generation capacity that emits carbon and consumes non-renewable fuel sources, but the power generation itself.

At some level of penetration, the system reaches the point where between baseline generators like nuclear and geothermal and the current level of wind generation is generating more power than is used in the daily demand trough. That's the initial economic target for cross-day storage ... storing that energy for sale during peak demand.

The next target is if there is either a determined effort to shift away from using natural gas for relatively quick starting power supply to mineral or biomass coal. Then cross-day storage can expand the peak-load generating capacity by firing up the thermal plant plant sooner, storing the power for delivery during peak demand.

However, as far as the location of those facilities, assuming the national long distance grid required to bring stranded wind from the high plains to consumers to east and west of them, that storage would most naturally be in facilities in the steepest terrain convenient to that grid.

On a side note, if reducing greenhouse gas emissions is a key rationale for the shift to sustainable, renewable sources of electricity, then its hard to work out why the wind ought to be used to displace natural gas, as opposed to using the wind to displace mineral coal. Thermal coal plants take longer to spin up from a standing start, so they would require a lot more stored power to cover their spin up than natural gas plants would do.

On a side note, if reducing greenhouse gas emissions is a key rationale for the shift to sustainable, renewable sources of electricity, then its hard to work out why the wind ought to be used to displace natural gas, as opposed to using the wind to displace mineral coal.

Because cost and security are also factors.  The Pickens plan is to use wind to displace natural gas, and then use the (domestic) gas to displace (expensive, imported) petroleum.  Petroleum also has a higher carbon impact than gas, though not as high as coal.

But if the target is to displace imported petroleum, electric long distance rail freight to displace imported petroleum through conservation seems a more direct approach. As is HSR to displace "short hop" airline traffic on 100mile to 500mile trips.

Of course, back up generation only consumes natural gas when it is doing the backing up, not when it is idle ... that is the point, after all, of preferring existing (sunk cost) or capital-equipment-efficient generating capacity be used for the back up generation. But it still seems as if there could be contention between gas turbine generation as back-up generating capacity and natural gas vehicles.

Ultimately, some form of bio-coal might be a natural complement as a back-up power source, given the ease of storing bio-coal ... am I correct in supposing that one appeal of direct carbon fuel cells for that kind of system is quicker start-up time than a thermal plant?

am I correct in supposing that one appeal of direct carbon fuel cells for that kind of system is quicker start-up time than a thermal plant?

Yes. The DCFC still has to be kept 'hot' due to elevated catalysis operation, but when insulated this is not a big parasitic load. The DCFC could throttle to full power in mere seconds. I don't know if it would operate well in partial load mode, but highly modular design is one of the benefits of DCFC and should allow high partial load flexibility simply by throttling a number of modules at peak efficiency while having the others on standby mode. If, for example, 50% load is required, and there are 100 modules, then 50 modules would run at max power (or more modules at slightly less than max power depending on the power vs efficiency curve) and the others would remain in standby. There have been difficulties with feeding and controlling of the grain size of the carbon particles into the cell plates, and issues with plate durability. But these look like manageable issues.

But if the target is to displace imported petroleum, electric long distance rail freight to displace imported petroleum through conservation seems a more direct approach.

Over-the-road freight traffic in the USA consumes far less fuel than light-duty vehicles (roughly 30% as much), and is a correspondingly less attractive target if the goal is either GHG reduction or displacing imported oil.  (If the goal is national security, it may tilt the other way.)  On the other hand, moving freight to rail (roughly 2/3 savings) and electrifying rail (to get the last 1/3) is very cost-effective for many routes without any change in fuel prices or technology.  That makes it time to move on it, regardless of what's done elsewhere.

am I correct in supposing that one appeal of direct carbon fuel cells for that kind of system is quicker start-up time than a thermal plant?

The appeal of DCFC is the combination of high efficiency, high turn-down ratio (a fuel cell can be throttled by a large factor easily, and almost instantly) and high temperature of exhaust heat (for e.g. industrial cogeneration).  At some throttling point I believe you have to shut off the air to halt the production of carbon monoxide, but modularity can address this easily as noted above.

Why can't they simply insulate the compressed air tank? The air gets hot during compression and stays hot. I also don't see efficiency as a big issue when we are wasting 15 cent kilowatt hours instead of four dollar gasoline. So it is 25% efficient. So what? Still cheaper than gas. The cost of infrastructure is probably the biggest impediment. We would need specialized air compressors all over.

My air conditioner is two stage. SEER 17.

Robert a Tucson

You would, but it's a little more complicated than that.

For various reasons, there is a limit on pressure ratios for compressor stages of about 2 (not absolute, especially in the case of screw compressors). If your storage volume is a cavern, you need several hundred psig discharge pressure; underground, a thousand psig; cylinders and receivers; up to several thousand psig. You need a multistage compressor to get that, with intercoolers between each stage. You'll get about 3/4 of your compression heat in the form of warm or hot water (anywhere from 100 deg F to 150 deg F, depending on the intercooler design).

Likewise when you use the compressed air you need a multi-stage expander, where the gas is heated between each stage. You can use burners to do that, or heat excahngers. If you saved the hot water you can use that. If you do use water HXs you get chilled water (40-50 deg F) as a byproduct.

I was thinking of storing the HOT air. Then when we expand, the temp drops back to room temperature. No heat exchanger or burner. But oh, we've stored a lot less air molecules. Our efficiency is better but we didn't store any more energy (or entropy as it were) for the same size and pressure rating tank.

Quoth robert2734:

I was thinking of storing the HOT air. Then when we expand, the temp drops back to room temperature.

This can be done, or the heat can be stored separately from the compressed air (a regenerator), increasing the mass of air storable in any given volume.  This would boost the efficiency of a CAES system substantially, and suitable heat-storage materials for moderate temperatures (e.g. crushed rock) are relatively cheap.

Quoth TJ:

For various reasons, there is a limit on pressure ratios for compressor stages of about 2

I don't think so.  Most heat-pump and A/C compressors are single-stage and have substantially higher pressure ratios.  The intercooled GE gas turbines have two stages, one intercooler and a pressure ratio of 42:1.

Likewise when you use the compressed air you need a multi-stage expander, where the gas is heated between each stage.

If you are burning fuel for the heat, it is far more efficient to add all the heat at the beginning.

Libelle's explanation of thermodynamics should have made this obvious.  For those of you who took calculus, dS = dQ/Tabs.  Adding a joule of heat at 300 K creates FOUR TIMES the entropy of a joule added at 1200 K; instead of heating from 300 K to 600 K three times, you should heat from 300 K to 1200 K once.  Every bit of entropy created is energy which must be rejected as waste heat, so minimizing entropy increase is the name of the game.

(Recycling "spent" heat back into the cycle is another way to avoid creating entropy.  This is why steam powerplants have "feedwater heaters" which take partially expanded steam and use it to pre-heat the water heading to the boiler.  Each bit of heat taken from partly cooled steam is heat that does not have to be taken from a much hotter flame in the boiler, so the increase in entropy is lower and more energy can be turned into useful work.)

Every bit of entropy created is energy which must be rejected as waste heat, so minimizing entropy increase is the name of the game.

That's the idea behind isothermal CAES. Impractical yes, but maybe now with carbon foam HX this could work at the MWe scale?

I don't think so. Most heat-pump and A/C compressors are single-stage and have substantially higher pressure ratios. The intercooled GE gas turbines have two stages, one intercooler and a pressure ratio of 42:1.

Small positive displacement compressors will pump to whatever discharge pressure you set, up until something breaks. Usually they're pretty close to isothermal also because they lose enought heat thru the cylinder walls. However, I've never seen a large centrifugal or axial stage on a stationary compressor with a ratio greater than around 2. Screw compressors yes.

If you're looking for high efficiency compressors running close to isothermal operation look at water-ring compressors. Nash compressors would be one vendor. They are limited in head to around 200 psig.

The drawback to putting all the heat in at the inlet of a multi-stage turboexpander is that the ICFM and the wheel diameter (or cylinder size if you go that route) is going to be really huge. A lot of investment and not much rangeability.

Usually [small positive displacement compressors are] pretty close to isothermal also because they lose enought heat thru the cylinder walls.

Funny, when I was working in climate control we expected substantial superheating of the vapor at the compressor outlet.  If the compression was isothermal, we would have had a saturated mixture.

However, I've never seen a large centrifugal or axial stage on a stationary compressor with a ratio greater than around 2.

By "stage", do you mean a single pair of rotor and stator rings?  It took me no time at all to find this piece on a centrifugal gas-turbine compressor aiming at 10:1 pressure ratio.

A system with a pressure ratio of 100:1 and intercooling after each 2:1 compression is going to require 6 intercoolers.  This is going to add a lot of cost and pressure loss.

The drawback to putting all the heat in at the inlet of a multi-stage turboexpander is that the ICFM and the wheel diameter (or cylinder size if you go that route) is going to be really huge.


All of today's best gas turbines do exactly what you say is a drawback:  add all the heat right at the beginning of a multi-stage turbo expander.  Since you're claiming to know more than the proven experts and contradicting easily-demonstrated principles of thermodynamics, it's obvious that you have no idea of what you're talking about.

Nice bluff, dude. Anyone who doesn't know what ICFM is (inlet cubic feet per minute) probably hasn't been anywhere near an actual compressor. But you tend to project lack of knowledge onto others to cover your own shortcomings.

Anyone who uses "cubic feet" rather than metric units isn't a self respecting global minded universal engineer.

Most interesting post and comments. It gives me some insight on a product made by Active Power that compresses air as part of a UPS system. Called CoolAir DC, they promote it for server installations where the cool air provided with decompression will (at least partially) substitute for the failed air conditioning that goes down with mains power. I never understood the function of the thermal storage unit in the system until I read your post. They have a block diagram at

I should add that it hasn't sold very well, though I don't know if that is due to technical issues or simply a lack of interest. The system appears quite simple and that usually translates into reliability.

The most obvious process for storing concentrating solar is to store the heat, using graphite or firebrick or the like.

The same process might appear useful here. For compressed air, using a very large volume of air to win against the square-cube issue, one might imagine, as appears to be considered above, storing the compressed air as hot rather than cold, perhaps using a little concentrating solar to top off against heat loss at the edges of the compressed storage volume. We are, after all, talking about a volume at 300 C, not 3000 C, so the challenge of building a container to take the temperature appears limited. Transmitting the hot compressed air any great distance to a central storage point might be the more difficult part of the operation, but compression only at the storage point perhaps solves this.

I don't spend much time looking here, so maybe somebody mentioned a simple and widely used storage system- hydraulic accumulaters, so called, in which a liquid is stored in a high pressure tank having a bubble of air stuffed with foamed material to soak up the heat of compression and thus eliminate all that loss of "heat of compression".

All this adds to a very efficient store and use cycle, in existence, and common. And good.

So, if you have already done it, fine, and if not, just look up hydraulic accumulator and get educated. I'm goin' back to bed.

PS. That's what I am doing for my own home power plant energy store.

Energy storage is going to be one of the most important topics in the coming years, because it does something amazing for the technologies that are becoming viable: wind and solar. The problem with these of course is that they are non-firm, which of course causes all sorts of problems for the grid.

Let me give an example; if you build a wind farm with say 100MW nameplate capacity, in some areas you will only get credit for 13% of that for the first few years until a known generation profile is developed for your farm. That's a huge derate and loss of profit. But with _certain_ types of storage, you can add a storage device sized about ~20% of the nameplate and provide roughly 98% firm power; giving you credit for nearly your full capacity. Storage does wonders for non-firm sources.

The AirCar uses a novel engine with an articulating con-ron to make it work. Ultimately air probably isn't the solution, although it works, and so do flywheels. Pumped hydro is great, and we have some spectacular units in the US like Bath County but the capital investment is very intense.

I think the real solution is actually some novel types of batteries; not the ones with good energy density, but ones with properties that make them good in terms of scalability and make MW sizes practical, and have lifetimes and designs that are reliable. And yes, these are starting to exist; with several competing technologies on the market. It gets efficiency very close to pumped hydro, which is probably about as good as you can hope for.

I'm not impressed by most of those bulk battery storage techs. They're either way too expensive for bulk storage, not efficient enough, or use large amounts of rare materials like vanadium (bad idea - vanadium is pretty rare and lots of vanadium on the market is a byproduct of purifying oil!). Or any combination of the above.

Underground pumped hydro and AACAES are some of the most promising bulk storage developments.

When I saw the 27% efficiency figure the thought came back to me of storing electricity as heat. A liquid metal in ceramic heating coil can reach temperatures well over 1000K. Take an underground cavity and fill it with pebbles. Put a electric heating coil on the bottom and loop of tubing on the top as a steam generator. Use the steam to generate electricity when needed. Could match or exceed the overall efficiency of a CAES system without the need for a sealed chamber.

Still not very good efficiency. 50% at best. Maybe in the future we'll have infrared nano antennas for decent efficiency...

One thing you never want to do is convert work directly to heat and then back to work. Do not stick a heat engine in your cycle if you don't have to. The second law of thermodynamics is not your friend.

The second law of thermodynamics is your friend! And Carnot is our buddy too! It says 99% heat engine efficiency is possible! Just not with any sane delta T and practical heat engine design :(

Seriously though, if infrared nano-antenna's pan out, then we may just see 70-80% round trip efficiency. Hopes for the future...

Two things to remember about electricity storage. It's never free and it always wastes. This article touches on these points for compressed air storage.

Here's a simple equation to calculate the unit output cost of electricity from any energy storage facility:

Cp = (P / e) + (T / M)

Cp = unit cost of output
P = unit price of electricity bought as input
e = efficiency of output to input
T = Output in units sold per year
M = annual fixed costs, ie the mortgage.

The first term is the excess of input to output while the second is how much it is used or amortization of capital costs.

The take-home point is that one is leveraged to prefer input sources that are both cheap and reliable. In other words, coal and nuclear are preferred partners for storage over solar and wind.

Nuclear doesn't store anything it just makes electricity continuously. Coal is burnt continuously in a steam turbine to make electricity can't store anything.

Natural gas and oil are excellent storage media.
Coal can be turned into synthesis gas which can be stored or turned into something(synthetic natural gas, hydrogen, methanol, CTL) which are storable but not without losing about 30% of the energy.

Actually solar is a perfectly reliable source of power in sunny climates and even moderately sunny climates but it doesn't make electricity at times when we need it. Same with large grid wind. These should the be basis of our electricity grid, not inflexible coal and nuclear which depend on limited fuel supplies.

We need to conserve the best forms of storage by putting a carbon tax on their use. Rather than worrying about storage, we should use excess electricity from renewables to make things we really do need like ammonia (fertilizer).

Right now we use natural gas to make ammonia--it's the most energy efficient way to do it(electrolysis costs requires about 30% more energy).
Instead what we should do, is to forget about expensive storing schemes for renewable electricity and use excess renewable electricity instead of natural gas to make ammonia at noon or on windy nights--saving the natural gas for our peaking energy needs. Now this will require a significant amount of renewable energy around 150 TWH(300 square miles of solar cells) but OTH the fuel will never run out. There are a range of commodities that can be made from electricity instead of gas.

The US uses something like 400 billion cubic feet per year of natural gas to make ammonia. OTH, the US uses 6000 billion cubic feet of gas for electricity generation but 7000 billion goes for industrial products( plastics, paper, steel, etc.).

The total cost of the electrolysers and ammonia synthesis equipment is very high, and running it at low capacity factors will compound that problem. If you're worried about expensive schemes, this is one thing we should not do.

I did explicitly state that this equation was for ELECTRICITY storage.

As to solar being "perfectly reliable", I guess if you can be fast and loose enough on your definition of "reliable" you might say that with a straight face. Saw a clip of Clinton with his infamous "depends on what your definition of 'is' is."

I'm not in the ammonia business so I'll leave it to you do explain the economics of your proposal. It might make more sense to import ammonia from gas-rich countries than to import LNG from the same places though.

Did you read about the farmers with $10 billion of yellowcake under their 200 acre farm in Virginia?

I'm not in the ammonia business so I'll leave it to you do explain the economics of your proposal.

I also doubt those economics.

But, since you're in the nuclear business, perhaps you could explain to us why new nuclear projects in the US are so darn expensive, and what are you going to do about it?


License application preparation - 18 months and $150 million
License review - 42 months and another $150 million or more
THEN you can start digging a hole - IF there is no political opposition, legal warfare, or change of heart in Washington.

60 year design life. 95% capacity factor goal.

One chance in a million for core damage and much less than that for overdosing the neighbors beyond occupational dose limits.

Tornado-proof, hurricane-proof, earthquake-proof, saboteur-proof, flood-proof.

20 year tax depreciation compared to 3 years for windmills (shorter is better for investors).

Lots of steel, lots of concrete, lots of high priced engineering and technical talent (such as your's truly.)

These machines are foundational infrastructure, built to last, to run through thick and thin, to survive the worst that nature and the evil of men can throw at them. They are the most powerful machines on the face of the earth. our designs and design goals are deeply conservative.

Could we come up with a design or designs that are cheaper?

Yes. The basic technology is 50 to 60 years old. It works well but we do have major advances on the drawing boards, like the pebble bed reactor. The prototype commercial unit is scheduled to begin construction in South Africa next year. I'd say, not bad for an industry that's fundamental scientific discovery (fission) happened in 1939.


I think the assertion for which I'd like to see evidence is that nuclear is cheaper than wind or solar.

The latest estimates for new nuclear seem to be reaching $6 per watt (average output). That's the same range that wind is in, and it looks like PV is getting there (First Solar's wholesale cost for panels is $1.12 per watt - with a 20% capacity factor we're in the same range). Given that nuclear costs about 2 $cents/KWH to run, and wind and solar are rather lower (a penny at most), I don't see evidence that nuclear is cheaper.

I realize intermittency is also a concern, but let's leave that aside for the moment (just for the sake of focus). Also, I'm not asserting that nuclear is more expensive, you have evidence for a description of nuclear as cheaper than wind or solar?

There is a LOT of confusion when people start quoting prices for ALL new generation. Quotes don't always mean the same thing.

First, what is the scope? Who owns the land that the source is built on, be it the 300 square miles of PV for the above poster's ammonia farm or the existing nuclear power plant site or the Altamont Pass? Who pays for any transmission line additions? Solar and geothermal in SoCal is getting the Sunrise transmission line built for them while a new nuke usually has to pay its own way. A hundred miles of dual circuit 500 kV lines costs a pretty penny.

Escalation assumptions are another area of fuzziness. It takes almost 10 years to bring a nuke on-line. What rate of wage inflation do we assume. What will be the price of a yard of concrete in 2018? How about the cost of money on those construction loans? I'll grant that getting a rooftop solar running in six months is an advantage but what will the same installation cost in 2018?

Tax treatment can be a deal killer or it can make a dumb idea a reality. Note I'm not a fan of production tax credits for any generation source, including nuclear. However, I can see that investors want some risk premium from government AGAINST government interference down the road. The biggest risk for nuclear is the wrong kind of government intervention like extensive licensing delays or design changes causing backfits. It is expensive to rip out major chunks of installed gear and replace it with the latest per government mandate - drives out schedules too.

Your impression that nuclear is $6,000/kW (the usual units) is a shock to me. We usually like to talk about $2,000/kW in overnight costs. That squeezes out the escalation uncertainty. Did a $600,000 house cost $200,000 in 1998? Many did. Your representation of wholesale solar costs also doesn't include power inverter/conditioner, installation, transmission connection (unless residential), etc etc.

We had a 1.4 MWe solar installation nearby at Cal State Hayward a couple of years ago. Taxes weren't an adder nor were equivalent land rents but yet the net cost per kW-hr was 55 cents excluding O&M. Even if the solar cells themselves were FREE, the attendant costs would still be substantial. And frankly, there is not that much inefficency or costs to be squeezed out of current PV cell design although solid state technology is not my field.

Guess I'll have to grant that the cost numbers are hard to calculate and agree to.

Solar and geothermal in SoCal is getting the Sunrise transmission line built for them while a new nuke usually has to pay its own way.

You're cherry picking your data. I could do that too; Pickens for example is building the largest wind farm in the US and is 100% paying for the transmission infrastructure. You want me to talk about Shoreham for a while? No, not representative.

Note I'm not a fan of production tax credits for any generation source, including nuclear. However, I can see that investors want some risk premium from government AGAINST government interference down the road.

Even after 80% gov't loan guarantees, private investment (real private investment, not some quasi goverment body) in nuclear projects is utterly dismal. Even Warren Buffet is pulling the plug.

Your impression that nuclear is $6,000/kW (the usual units) is a shock to me. We usually like to talk about $2,000/kW in overnight costs. That squeezes out the escalation uncertainty.

Even then that's way too low for overnight cost based on projects being commissioned/started in the US. It's not entirely clear if this is first of a kind cost premium (development cost etc). I suspect it has more to do with a comatized industry waking up and finding it's got the shakes. Industrial capacity is moribund. This won't change rapidly due to the lack of private investment. It's quite a fix really.

The figures I've seen are all in the 5000-8000 per kWe total project cost (all-in plus infrastructure). For wind it looks like somewhere in the 2000-3000 per kWe range (all-in plus infrastructure). These estimates are from the people who are building it and paying for it. Surely you agree that total cost is what we should be comparing? After all, if we don't do that, diesel generators are actually cheapest, with overnight costs less than 200 per kW. So should we build more diesel generators? Levelized cost metric deals with this pitfall. So if we're only going to use one metric, then use levelized cost.

License application preparation - 18 months and $150 million
License review - 42 months and another $150 million or more

That's 300 million. Typical 1 GWe power station, that's maybe 30 cents a Watt. Now explain where the other 5000 to 8000 million are going to. That's 5 to 8 bucks a Watt.

60 year design life. 95% capacity factor goal.

Using 40 years and 85% capacity factor is less risky for fleet performance. The reliability of new nuclear technology is unknown; if you don't do sensitivity analyis, it ain't worth crap.

Could we come up with a design or designs that are cheaper?

Yes. The basic technology is 50 to 60 years old. It works well but we do have major advances on the drawing boards, like the pebble bed reactor. The prototype commercial unit is scheduled to begin construction in South Africa next year. I'd say, not bad for an industry that's fundamental scientific discovery (fission) happened in 1939.

Not bad, but developments aren't nearly as impressive as wind's innovation cycle, which has become slightly cheaper than nuclear in a much shorter development time. Considering it's swift innovation cycle, it's unlikely that wind will lose the advantage. WInd is here now, and it's going to be big. For nuclear power there's a big questionmark, which means liability for the future. Not if we pursue more nuclear power, but yes if we do what many (but not all) nuclear enthusiasts would like - nuclear absolutism. The way things are looking now, I don't see a plausible reason why wind should be dropped in favor of nuclear power, policy wise.

The 60 year life and 95% capacity factor are in the EPC contract draft I'm reviewing (EPC = engineering, procurement, construction). The existing fleet is running 90% so it just takes designing for shorter outages and 24 month refueling cycles to reach that goal. The 60 year life has been assured for existing plants and is designed into new ones.

I think you've avoided my point of what will the alternatives cost when a nuke starts up. We're looking pretty far into the future with these long lead projects. What will your finances look like in 8 years?

Still, planners for electric infrastructure MUST work under uncertainty. With 34 new reactor applications either in the hopper or promised a review slot, Big Money is making a big investment in nuclear, even if Warren Buffett thinks his project is a no-go.

We have to remember compound interest. That $150 million at t=0 faces compounding at 10% per annum. After 10 years, that adds up. Timing of cash flows is a big part of project analysis. Lots of squabbling in contract negotiations revolve around when cash changes hands.

Wind does have a niche so long as natural gas price is substantial. It is worth the fuel displacement costs for turning down CCGT plants. That's pretty much the nut of T. Boone Pickens' argument. Beyond that and I think it a waste of resources.

Don't get me started on residential solar net meeting.

The 60 year life and 95% capacity factor are in the EPC contract draft I'm reviewing (EPC = engineering, procurement, construction). The existing fleet is running 90% so it just takes designing for shorter outages and 24 month refueling cycles to reach that goal. The 60 year life has been assured for existing plants and is designed into new ones.

I am not aware of new nuclear projects being contracted for 60 year PPA in the US. Oh, and where's the empirical evidence for 60 year fleet life? Design is one thing. Proof is another. Even the best designers don't know everything.

I think you've avoided my point of what will the alternatives cost when a nuke starts up. We're looking pretty far into the future with these long lead projects. What will your finances look like in 8 years?

I support new nuclear projects. I think you've avoided my point that wind is here now and it's growing rapidly. Nuclear isn't growing fast enough for it to be relied on as silver bullet, and the industrial and financing aspects retard short term exponential growth. For the next 8 years, don't expect exponential nuclear growth. What often confuses nuclear enthusiasts are the large differences between the nuclear generating industries and the nuclear construction industries.

Still, planners for electric infrastructure MUST work under uncertainty. With 34 new reactor applications either in the hopper or promised a review slot, Big Money is making a big investment in nuclear, even if Warren Buffett thinks his project is a no-go.

Central planners are one of the root causes of the opaque situation that is nuclear power economics; planners say build nukes, the market says no, or nothing (which can often be interpreted as no). Big money? Big government money mostly, yes. There's also big government money behind ITER and the war in Iraq.

Wind does have a niche so long as natural gas price is substantial. It is worth the fuel displacement costs for turning down CCGT plants. That's pretty much the nut of T. Boone Pickens' argument. Beyond that and I think it a waste of resources.

I would judge an investment a waste of resources if the oppertunity costs are substantially higher than for alternatives. This appears to be the case for nuclear power right now, but it may change in the future so a strategical policy would direct some incentives towards nuclear, mostly towards R&D and efforts to reduce the oppertunity costs (in particular capital) and reduce project uncertainties. I believe it's a bad thing to directly subsidize projects. A small per kWh subsidy makes some sense, so I disagree with you on PTCs, as long as they're not too big (less than 5 cents), they don't create inefficiencies in markets.

Central planners are one of the root causes of the opaque situation that is nuclear power economics; planners say build nukes, the market says no, or nothing (which can often be interpreted as no).

Jerome Guillet has written on this subject, and noted that market uncertainties lead investors to minimize sunk costs even if this leads to highly sub-optimal results.  Government can fix this by legislating certainty (e.g. feed-in tariffs) and using its borrowing power to reduce uncertainty of interest rates.

I agree with that, I'm an advocate of 'enhanced markets'. Feed-in tariffs work. One thing to be careful about is not to make the tariff absurdly high (like 50 cents per kWhe) as that actually promotes inefficient uncompetitive solutions. Another is to make it equitable: all low carbon sources should receive it IMHO, otherwise the central planners favoritism gets in the way again and the problem of opacity is still unsolved. In an equitable feed-in tariff market, utilities and investors will still decide what's most practical/desirable.

The marginal cost of wind power is so low that in an off-peak setting it rules the roost ... if wind capacity is expended substantially, off-peak wind will certainly be cheaper than coal.

Of course, under full economic cost, coal would lose out to a high quality wind resource under any circumstances, and even under the current regime of cost shifting onto third parties, the commercial cost of coal has been rising.

So under the formula above,

Average Cost = Average Variable Cost + Average Fixed Cost

... the cheapest AVC is from surplus wind.

(Obviously the availability of surplus wind is directly related to, and the mean and variance inversely related to, the degree to which different wind resource regions are connected into a single long distance grid.)

Now, if at some times there is not surplus wind, so that the marginal cost producer is nuclear or coal, then in that case, the storage capacity can be charged the off-peak power at that cost.

So with the storage as a network capability rather than dedicated to specific power producers, that is the same AFC and a lower AVC for a network with sufficient wind generating capacity to carry a substantial average share of daytime demand, since that network will have some periods of surplus off-peak power generation.

The marginal cost of wind is indeed next to nothing. that's why it is usually "must run" in economic dispatch orders. And given it's lack of reliability, why would one ever have a surplus?

If you want to charge the electric storage facility only marginal cost for wind-produced juice, who pays for the fixed costs? And wouldn't everyone else want that next-to-nothing electricity?

In a comparison with the status quo, then given the full economic cost of coal, you'd harvest wind power because its the cheaper source of electricity, and use more expensive power like coal and natural gas when the wind power is not there.

In order to harvest that power, there has to be a rate structure in place that covers the average cost of the wind turbines (or other equipment to harvest volatile renewable power sources).

That will be paid for by people demanding electricity at the time that the power is generated.

Given that rate structure, and given that there is no substantial operating cost to harvest the power as it is available, there will be periods in the day when there will be a surplus available for harvesting over the amount of electricity currently demanded.

That creates the opportunity to store electricity in periods of surplus production for sale later that same day during peak demand periods.

And that is a network capacity rather than a production-facility capacity, because there is no technical need for the facility harvesting the volatile renewable energy source to have a storage capacity. Its purely opportunistic.

Of course, there are rate regimes in which facility level storage is a critical part of harvesting a volatile power source, but the rate regime in force is a policy choice, not a law of nature.

Exelent idea:

Sink a massive cason (or however you spell it) deep into the ocean with large water turbines at the bottom that use the pressure of the incoming water to generate electricity. When electricity is in excess capacity, the pumps are sealed and water is then pumped out of the top of the cason.

Any thoughts?

This cason would be truly massive. A more straighforward way to use the ocean is to dam a large and deep bay and pump the water in and out.

Another idea I have been contemplating is to build an artificial "lake" into the ocean, by sinking massive concrete walls encircling a large area and pumping the sea water in and out. For example a internal "lake" of 10km in diameter at a depth of 200 meters would contain 4,361 GWh of energy. For comparison the daily consumption of electricity in US is about 11,000 GWh so a single one of those could theoretically balance the whole US electricity grid. I don't know what would be the cost of such project though, probably it would need to be devided in many smaller projects to make sense.

Take a look at the Ludington, MI. pumped storage facility - it's exactly what you describe, on Lake Michigan. It's been operating for 30+ years.

I checked it out here, and it appears to me that Ludington is a conventional pumped storage with lower and upper reservoir.

In my proposal there is no "upper" reservoir, the sea itself is used as an upper reservoir (or the lake, but it must be a very deep one). Here is a quick schema of what I imagine:

Image Hosted by

The problem I see is with building those massive walls, but I don't think it is unsurmountable.

Well, Ludington uses Lake Michigan for the lower reservoir. I'm not sure I see the benefit of putting one's reservoir inside the lake or ocean: I think it would be much easier to build those walls on dry land. Probably easiest and cheapest to do what they did at Ludington: get some elevation from digging, and some from building a wall.

The benefit is that you don't need to do any digging to implement this setup. The main problem of pumped storage is that there are few suitable sites with big enough denivelation, capable of hosting the upper reservoir. This idea addresses this problem - all we need would be deep ocean waters near shore and near demand centers - and there are thousands of square miles of ocean that fit in - more than enough for all practical purposes.

It will require detailed engineering to see whether it is feasible though.

It looks good to me. Its concrete would be loaded in compression rather than in tension, as the walls of a giant beaker would be if the water level were higher inside than outside.

Perhaps it should not depend on the wall being joined to the seabed at the bottom. If it had its own bottom, it could be built in water deeper than 200 m.

Perhaps there is a way segments could sit neutrally buoyant in the sea and be driven together. Each segment would be L-shaped as seen from the side, pie-slice-shaped as seen from above. Once a little water was pumped out of the region they would, when driven together, collectively enclose, they'd be held strongly together from then on. The detailed engineering would include getting them to sit in the right orientation before they were put together.

--- G.R.L. Cowan, H2 energy fan 'til ~1996

Took you long enough to figure out that hydrogen transportation sucks!


Took you long enough to figure out that hydrogen transportation sucks!


When confined in an energy reservoir system of finite volume, yes, hydrogen exerts pressure that rises asymptotically towards zero.

So I didn't figure it out until 1996, or maybe even 1998. Think I might have been distracted. But I reported my findings fairly promptly. How could it be, then, that ten years later, a whole journal would exist, dedicated to how the nuclear people can get in on the ground floor of the same ride they were taken for in 1970.

--- G.R.L. Cowan, H2 energy fan 'til ~1996

You're quite a character, Graham!

I think the problem lies mainly with the inability of most people to understand the macro-economical (and energetical) implications of entropy retarded efficiency in a finite world. What works on a tiny scale, does not always work on a scale large enough to really matter. And why should it, if alternatives that are better in every aspect already exist? Favoritism?

I think there won't be any problem to make those segments fit in each other by putting guiding channels and guiders in the odd and even ones (I'm not familiar with the exact engineering terms).

I also thought of onshore reservoir before posting this idea, but I was concerned of the land requirements and the failure mode of one like that. I also thought that in ocean it will be more easy to build it by lowering down the pre-built segments from a barge, but I'm not 100% certain about this one.

No practical use for this on a large scale but a cool Thermodynamic oddity.

Vortex Tubes

High pressure air in: Cooled air out one side, hot air out the other side.

One look at the coefficient of performance (and the fraction of Carnot efficiency) tells you why these things are oddities:  they make the Plymouth Roadrunner Super Bird look efficient.

Oh man, instead of electric cars now people are starting to talk about compressed air for automotive travel.
Now the Polyanna's are really grasping at straws.


So...I believe a good physics lesson on TV is in order for the public instead of watching Friends or whatever the hell the masses are watching these days.

BTW, do not let the current price of a barrel of oil fool you, the economy and personal travel are way down - perhaps by 50% for nonessential travel where I live - that is the only reason the price of a barrel has gone down...somewhat?!

Big events are just around the corner; within a year a gallon of gas will cost between $8 - $10 in the US and most will not know the reason why. They will figure it is because of war in the Middle East - the way they always spin s#*t on TV.It seems these days war and economic downturns are concomitant in time with one another.Remember the dotcom bust and 911 - how convenient.There will not be a lag time between the current economic bust and war the way there was in the 1930's, it is going to transpire very rapidly.
Time to start reading your Bibles.

The efficiency problems with compressing or expanding air can be solved involving liquids providing a heat exchange during compression or expansion. Three basic systems can be used, and we tested them all successfully:
1) mist injection with any kind of piston
2) plate immersion (works only with liquid pistons)
3) embedded bubbles in water columns
-The injection method is the most effective for compression but difficult for expansion: you will find an exhaustive description thereof by M. Saunders in the Scientific American Supplement Vol. XXXI, No. 799, 1891; it is used recently for highest efficiency H2 compression up to 700 bar under a DOE contract with special liquids (contact: Maryann.Daniel Injecting Water limits pressure as increasing solution of air in the mist occurs due to huge interchange area.
-The immersion method (see Mr. Knith’s Patent 586'000 / 1897) has less contact problems with the air, as only the piston ective area is exposed. It is also well suited for expansion, but for pressures up to 3000 PSI special liquids should be used.
-The embedded bubbles or direct hydraulic compression is the oldest way of generating compressed air without moving parts, used under the name of “trompe” in forges etc. Modern direct hydraulic compressors known as „Taylor compressor“ are still in use, reports about a former installation at Victoria Copper Mine near Rockland, Ontonagon County, Michigan gives an extraction of almost 2000 HP with an efficiency of 82%. (see H.J.Torkelson: “Air Compression and Transmission“, McGraw-Hill 1913). Note that the expansion operation of the embedded bubble machine is the „Mammooth Pump“, widely used for abrasive water. This direct power extraction even from otherwise useless small waterfalls without moving parts led many authors to consider compressed air as the only genuine combustion-free power source from nature with inherent storage capacity, unlike electricity or methane.

So, where is the problem??