Compressed Air Energy Storage

I have always been a big fan of wind power. But one of the knocks on wind is that it is intermittent. Since electrical demand probably won’t match up very well with wind fluctuations, installed wind capacity does not displace conventional power generation in a 1 to 1 ratio. For example, I have seen it claimed that 2,000 megawatts of installed wind energy still requires 1,800 megawatts of standby power for when the wind isn’t blowing. (1)

Clearly a storage system is needed. During times of high wind flow and low demand, the excess energy could be stored in something akin to a giant battery. When the wind isn’t blowing, users would pull from the “battery”. I have given a lot of thought over the past couple of years as to just what form such a storage system would take. I could envision several different options. One, air could be compressed into a storage system and then discharged through a turbine as needed. Two, water could be pumped uphill, and then be allowed to flow back through a turbine as needed. Three, water could be split to hydrogen and oxygen. I don’t like this option as much, because most electrolysis is inefficient and hydrogen storage is problematic.

(Incidentally, a variation of this third option was recently brought to my attention, in which excess wind power is used to make methanol, ethanol, or even ammonium nitrate fertilizer. For an excellent comprehensive overview of this option, combined with an impressive overview of wind energy potential in the Great Lakes area, see David Bradley’s report A Great Potential: The Great Lakes as a Regional Renewable Energy Source.)

Imagine my surprise this weekend to learn that while I have been daydreaming about a wind energy storage system, someone is in the process of doing it. Furthermore, others have previously blogged on it. I felt a bit like Rip van Winkle after waking up from his long nap. How could I have missed such an important development? The storage system is called compressed air energy storage (CAES). A January article from MSNBC explains the concept in Store wind power for later use? Cities bet on it:

A group of Iowa cities intends to not only harness the wind, but also capture it, store it underground and use it to help make electricity when demand peaks.

Members of the Iowa Association of Municipal Utilities have invested in a proposed power plant that would use wind turbines to drive compressed air into underground aquifers. The air would be released to generate electricity when needed.

The plant will use power from its own wind turbines, supplemented by cheaper electricity bought at off-peak times, to force air into rock formations at least 2,000 feet underground.

Current plans call for pressurized storage of tens of billions of cubic feet of air in rock formations deep underground. (2)

If you think I was surprised by that, imagine my surprise upon reading this from the same article:

Only two other underground compressed air plants are in operation. A plant in Huntorf, Germany, was built more than 23 years ago and a plant in McIntosh, Ala., is 11 years old. Both store compressed air in underground salt caverns.

Iowa’s project is unique in that it would use wind power to store the air and combine it with massive underground storage capacity.

The Germany and Alabama plants store hundreds of thousands of cubic feet of air in a thermos-bottle shaped container installed in the salt mines. The Iowa project would use naturally occurring pockets embedded in sand or sandstone formations sealed by shale or other rock.

So, a plant in Alabama has been using compressed air storage successfully for 11 years, and I didn’t know about it until this weekend. The only difference is that they aren’t using wind to do it. The Iowa plant will be the first to do that, but others will probably follow.

To be sure, such a storage system is not universally applicable. You need some kind of large, airtight, underground cavern. There are a lot of these in the United States, but they need to be located near a source of wind. Although, now that I think about it, I see no reason such a system couldn’t also be paired with solar or tidal generation systems, storing their excess energy using the same concept.

The plant is scheduled to come online in 2010. I wish them great success, and look forward to hearing reports after they start up.


1. “Airtricity’s rise and rise leaves criticism blowing in the wind”, Irish Examiner, April 30, 2005.

2. “Store wind power for later use? Cities bet on it”,, January 4, 2006.

24 thoughts on “Compressed Air Energy Storage”

  1. An auto company in Europe is also working on a short-range, commuter auto that uses compressed air instead of electricity or fossil fuels.

    The engine is powered by compressed air from a storage tank, and when braking and slowing down, the proces is reversed and air is compressed and forced back into the storage tank.

    Obviously no pollution except at the point where the electricity is generated that compresses the air. Get that electricity from a wind turbine/compressed air storage sytem and there would be no pollution at all.

  2. The issue with CAES is that it isn’t a pure storage system. When a gas expands it will cool; in this case it goes well below freezing. Needless to say having ice form on your turboexpander blades is a bad idea.

    In a CAES system the pressure is used to provide air to a natural gas turbine instead of using a compressor. Since a compressor might consume 40 % of the energy in a turbine that fraction can use stored power from the pressurized air in the salt cavern.

    A typically CAES system has a round-trip efficiency in the 73 % range, which is below vanadium flow batteries, pumped storage and of course lithium-ion batteries (but they aren’t utility scale). It’s about on par with most flywheel schemes.

  3. Needless to say having ice form on your turboexpander blades is a bad idea.

    I saw a graphic of the process this weekend, and if I recall correctly they were condensing the water before they injected it into the caverns.

    A typically CAES system has a round-trip efficiency in the 73 % range, which is below vanadium flow batteries, pumped storage and of course lithium-ion batteries (but they aren’t utility scale).

    One source reported that the Alabama facility is 82% efficient, but I never saw any technical information to back this claim up. In researching this, I saw vanadium flow batteries mentioned several times. I don’t know anything about them. I need to look into that.


  4. There is a much simpler solution which avoids the necessity for costly storage plants… just dump the extra power on to the grid, and let the market deal with it.

    However, for this to work, electrical devices have to be “intelligent” and capable of responding to electrical prices dynamically. For example, there is no reason why my refridgerator has to run at peak times — as long as it is within a certain tolerance band of the “ideal” temperature, the refridgerator can afford to wait for lower electrical prices by projecting the temperature bleed and optimizing electrical consumption vs. the need to stay close to ideal temperatures.

    When wind power becomes available, it just dumps onto the grid, thereby causing electricity prices to fall. Intelligent consuming devices then respond to lower power prices by consuming electricity.

    The solution to random supply is simply “flexible” demand… as far as I know, there are some experimental devices already in the works (eg: washers/dryers, I believe), but this is, of course, many years off…

  5. “..seen it claimed that 2,000 megawatts of installed wind energy still requires 1,800 megawatts of standby power for when the wind isn’t blowing.”

    This is the second time I’ve seen wind described this way (the first by a small German transmission system operator). It seems to me to be a clear sign of prejudice against wind, as it is a very misleading way to describe wind’s contribution to system capacity.

    Wind generators have a rated capacity (e.g., 3 Megawatts) which is designed to be substantially higher than the normal operating range, in order to maximize energy capture. A reasonably well placed wind turbine might have a capacity of 3 MW, and utilization of 33%, or an expected average production of 1 MW. A coal or nuclear plant would be considered to be badly failing it’s design goals at 33%, but a wind farm very likely would be considered quite successful.

    This means that the maximum contribution to peak capacity expected from such a wind farm would be 33% of capacity. You could describe a 2GW wind farm as requiring backup capacity of 1.33 GW, and you would be describing a successful system. It wouldn’t sound like it, though.

    A 2 GW system that required 1.8 GW backup would be providing .2 GW capacity contribution, or 10%. That would be roughly 1/3 of what would be ideally expected. On the other hand, the same Irish TSO source said that a smaller 1 GW system would provide a 20% capacity contribution, which would be pretty good at about 2/3 of the ideal.

    You have to keep in mind that this is a very small system – we’re talking about a 6.5 GW (peak)system, compared to 905 GW in the US. Ireland is very small, and only barely interconnected to the UK (about 6% of capacity). This means that you would expect them to have a relatively hard time coping with 2 GW of wind capacity.

  6. What about safety? If you have a lot of air at high pressure, and the storage tank fails then all the stored energy gets released instantaneously.

    I know scuba tanks (at 3000 psi) need to be handled with care and inspected regularly. What pressure does it take to power a car? Now what happens if the storage system is compromised in a serious accident?

  7. Compressed hydrogen at high pressure would be much more dangerous. However, tests have confirmed the safety of hydrogen tanks:

    Many real-life tests have demonstrated the safety of pressurized hydrogen storage. Simulated 55 mph crash tests left the car totaled, but the hydrogen tank intact. To prove the safety of its hydrogen vehicles, BMW tested its hydrogen tanks in a series of accident simulations that included collision, fire and tank ruptures. In all cases, the hydrogen cars fared as well as conventional gasoline vehicles. And hydrogen-fueled cars are designed to preclude the possibility of leaked hydrogen collecting within the vehicle.

    From: Is Hydrogen Dangerous?


  8. Robert;
    I’ve been toying with another form of stored energy, related to pumped water storage, which is simply lifting dead-weights. This seems at first sillier than water-lifting or compressed air, but spares the losses associated with moving fluids; viscosity, turbulence and cohesion. I don’t know if the direct lifting using chain or cable (pneumatic gets back into compressors again)

    Initially, the picture in my head was to use towers (and the thought was for home-scale, not utility.) but then clearly, the safety and engineering potential of digging shafts looks more likely. A catastrophic collapse would be fairly tame, unless you were working in the bottom of the shaft.

    I would think multiple shafts with maybe 15-20 tonne loads would tie together for a very managable release, as wind or solar supplies varied.

    Instead of lifting them with transferred electric, I would think they’d get their own turbines, designed to lift with direct, mechanical connections. Escapements can be built that allow for the generated output (dropping) to be managed independently of the wind as it lifts the weights, as well, making for a more ‘on demand’ system.

  9. …designed to lift with direct, mechanical connections. Escapements can be built that allow for the generated output (dropping) to be managed independently

    Good thinking — and there is even a precedent:

    I have a kuckoo clock that works exactly the same way, only on a much smaller scale.

    It would be much like pumping water uphill into a reservoir or holding tank, only instead of liquid with weight, it would be a solid mass. Use a solid mass with a very high density such as depleted uranium, and the towers could be shorter — or the shafts less deep.

  10. Nigel;
    Right. If your load is carried under it’s own pulley, for instance, then one end of the cable rising from it goes to the lifting mechanism, and the other end to the generator, with spare managed between those units to complete a loop of the load-bearing material, which advances throughout the process.

    I’m trusting the Kuckoo Clock example wasn’t really the buried review of this idea, with a little DU frosting on top. I’m so slow to pick up sarcasm in text.. but I can’t see that this idea is really worse than either of the pumped systems mentioned. Like most storage formats, it’s heavy on investment, but uses a constant in gravity, depending on nothing but sheer mass (waste materials..) to function repeatedly. and we’ve gotta do something with that Uranium..

  11. I’m trusting the Kuckoo Clock example wasn’t really the buried review of this idea, with a little DU frosting on top. I’m so slow to pick up sarcasm in text…


    No sarcasm. I think your idea has potential, and I wish I had thought of it.

    Not just for storing energy in wind farms, but also for storing energy at individual houses and small businesses.

    At a house or small business you might have a tower or tripod arrangement that would allow a weight of a few tons to be lifted 50 or 60 feet. You could use electricity at off-hour rates at oh-dark-thirty to lift the weight, and then use the descending weight during the day to power an alternator or generator to run your household or business.

    Over the long-run it should be more efficient than a battery since there would be no chemical deterioration. My kuckoo clock has been chugging along for the last 35 years and I’ve never had to change a battery — just pull those weights back to the top each day.

    This would be a simple and elegant way to store energy for later use.

    I commend your vision, and really am curious how it would scale up. How tall a tower and how much weight would you need to power a typical house?


    Nigel Gamecock

  12. Any word on what the efficiency of a compressed air storage mechanism is? You’d have losses while compressing, and losses again when decompressiong through a turbine. My guess is that this technique would be very inefficient and relatively expensive.

    Someone suggested lifting weights and storing the energy that way, and someone else suggested lifting water. The water technique is called pumped hydro, and can be relatively efficient.

    On a residential basis, lets say that the house has a 20m water tower, to store enough energy to run the house for half a day (about 12kwh), youd need to have a capacity of around 200000 litres, which is about the volume of a small swimming pool.

  13. I am not sure the salt cavern needs to be near the wind (or solar), but rather I think it needs to be roughly on the power lines between the wind farm and the people who use the power.
    On a related note – can you take west Texas wind power, and use it to fill the natural gas salt caverns with hydrogen, and then simply pump the hydrogen into the exsiting Natural gas pipeline network? Obviously the losses would be higher, but not using the hydrogen to regenerate electricity reduces these losses somewhat. It is also nice having a gas heater when the electricity is out.

  14. In areas without hills but with a low water table (e.g., 300 feet down), it seems windmill energy could be stored by pumping water up into a pond when the wind blows, then letting it run back down the well through a submersible generator when power is needed.

  15. An Aggie in Scotland? You’re far from home. I am an engineer doing village level energy systems in developing countries. Have you heard of any systems using compressed air storage on a small (<50kW) scale?

  16. The problem with home based mechanical storage systems is sheer size and expense. Say for example compressed air, or a hydro project that pumps water into a pond at higher elevations. The average home load for a single family home is about 2kW, but the peak load might be 10 kW. That means for most of the day your load is less than 2kW but when you come home and run your appliances you need lots of power quickly. Up to 20kWhrs of power over the course of a few hours.

    20kWhrs of power is a lot for something to be stored mechanically in your back yard.

    Example: a weight in a shaft. Stored energy (assume 100% efficiency) = gravity x mass x height of drop.

    20kWhrs = 72000000 Joules of energy

    assume you have a 1000kG mass (roughly the weight of a small car)
    youll need a shaft or tower with a7.3 Kilometer drop!

    Its easier to store energy mechanically if you have a naturally occuring or pre exsisting underground cavern, or a massive lake to use. The difference between peak and average load is also not as severe for utilty companies because the load is shared over manu users. But on a home scale mechanical storage with large differences between peak and average load becomes unfeasible from a cost and scale point of view. It would actually be easier to simply oversize your generating equipment – and diversify it (solar+wind+hydro combined) than it would be to try to shrink it down near your average load and store off peak load.

    The cheapest of all solutions is to shrink your load by use of energy efficient alternatives – but most people prefer to think of solutions that dont involve changing their lifestyles.

    Which ever way you go do the math and sadly you end up with massive scale projects for a home based storage system. Massive storage tanks bigger than your house with many 1000’s of PSI pressure would be needed for a CAES system, huge ponds with 100s of feet elevation, or massive weights. Even flywheels would need to be huge to store 20kWhrs. This is exactly why batteries and their nasty chemicals came into being in the 1st place. Amazing amounts of energy can be stored in such a small box, its hard to beat the energy storage density of batterys.

Comments are closed.