Useful questions re: CAES

Here is what I want to understand about CAES. When you compress air it gets hot. When you let it out of a vessel where it has been stored at pressure it gets cold.

It seems to me that much energy will be lost if that heat is not recovered or if the cold is just allowed to dissipate.

Another issue is how much will these things cost?
 

This study estimated $504/kW for the CAES system.  Not sure what this would be for a system designed to substitute equipment for co-fuel.

Yes, much of the heat of compression would be lost.  This can be reduced using intercooling between compression stages, but it will be significant regardless.  A concept I have not had time to analyze (or look for existing analyses) is to inject water into the compression system and capture the heat energy as steam mixed with the air.  This will reduce the temperature rise to a managable level, allowing the compressed air/vapor to be stored without discarding heat.  Most of the water would be recovered as condensate during the expansion phase.
 

CAES achieves its high efficiency (~ 70 %) because it's only used to bypass the compressor on a natural gas turbine. If you tried to run compressed gas to directly run a turbine you would take huge losses from the compressibility of the air and have issues with ice formation, etc. I think you would be lucky to get 35 % round-trip efficiency.

The actual stage of compressing the gas is, of course, adiabatic. I don't know if you could realistically build a regenerator for such a system.
 

"CAES achieves its high efficiency (~ 70 %) because it's only used to bypass the compressor on a natural gas turbine."

The fuel efficiency claimed for some of these things is ~80% (ignoring electric input, which is presumed to come from wind and may represent generation that could not be used due to transmission limitations).

"I think you would be lucky to get 35 % round-trip efficiency."

If you go through the numbers, the overall efficiency (output divided by all inputs) of the latest CAES concepts is in the region of 50%.

"I don't know if you could realistically build a regenerator for such a system."

Regenerator?  Interesting idea.  Suppose you could fill a large chamber with crushed rock, density 3.5 g/cc and 30% void space.  The heat capacity of rock is about 0.84 kJ/kg-K, so a spherical chamber 100 meters in diameter would hold 1.28 billion kg of rock with a total heat capacity of 1.08 TJ/K.  Hot air gets pumped in through a pipe to the center, and filters around rocks until it leaves via ports at the edge; this is reversed when drawing air out.  If you assume that the hot air going in is at 600°C and comes out at 200°C and the useful range on the reverse goes from 600°C down to 400°C, the energy capacity of the regenerator would be somewhere between 200 TJ and 400 TJ.  That's 56-111 gigawatt-hours, a non-trivial quantity for such a small volume (especially if you mined it out underground and it required no surface space).

Regenerators could be placed between compression and expansion stages, so a system could use more than one.  It looks like they merit further investigation.
 

How Compressed Air Could Power the Future by Michael Schirber for LiveScience on 04 June 2008:

Wind power is unreliable. No one can turn up the wind every time electricity demand peaks. So some utilities are looking at ways to bottle up the wind's energy and store it underground for later use. ... Compressed air energy storage (CAES) uses off-peak electricity from wind farms or other sources to pump air underground. ... However, a good portion of the input energy is lost in this process, making CAES one of the least efficient storage technologies available. ...

Storing off-peak electricity is not new. By far the most common method is to pump water up to an elevated reservoir and then release it to drive an electric generator when demand calls for it. Once "charged," these pumped hydro systems — of which around 300 exist worldwide — can supply 1,000 megawatts of power for several hours. However, there are few places left with available water and the right topography, Marquardt said.

CAES can supply around 100 megawatts of power for several hours, and the needed geological formations (abandoned mines, salt caverns, aquifers) can be found around the world. Other storage devices, such as batteries and flywheels, cannot store nearly the same amount of energy and are much more expensive to install than CAES.

Currently only two operating CAES facilities exist in Germany and Alabama. They each use salt caverns with several hundreds of thousands of cubic meters of space (roughly the volume of 100 Olympic-sized swimming pools). Using off-peak electricity, air is compressed to around 1,000 psi (or 70 times atmospheric pressure), which raises its temperature to more than 600 degrees Celsius (1,100 degrees Fahrenheit). This is far too hot to pump underground, so the air is cooled to about 50 degrees Celsius (120 degrees Fahrenheit). Unfortunately, the air must be reheated on release in order to turn a turbine. This extra reheating energy (usually provided by burning natural gas) means CAES has a relatively low efficiency of about 50 percent: for every kilowatt-hour of energy going in, only 0.5 kilowatt-hour of energy can be taken out.

* * *

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. ... The efficiency could be increased to 70 percent ...

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.

Once the technical difficulties are all ironed out, Marquardt expects a first demonstrator project — supplying around 30 megawatts of power — sometime around 2012. The future goal is to have an AA-CAES facility that can supply 10 times that. ...
 

It seems to me that the better way would be to use a big pool of watter (suitably insulated) as the heat/cold resivoir.
 

Not sure if water would work all that well; it would require a sealed storage space, and the temperature range is limited by the strength of the rock (to contain the vapor pressure).  This is also the problem my mixed air/steam idea has; storage media like sandstones would condense the water and not return the heat until a very large volume of rock had been warmed considerably.

Heat storage is definitely the way to go, through.
 

If the purpose of heating the gas is to avoid ice forming on the turbine blades, then high grade heat shouldn't be necessary, right? (as in, to accomplish high thermodynamic efficiency in a heat engine).

Cost would be important, and low temperature heat storage means more volume, insulation and thermal mass required, thus increasing the costs. But that's not a problem if water is used - cheap, especially considering the low pressure (lower than 100 degrees C) which means an inexpensive resevoir, like simple, huge thinwalled tanks aboveground, covered with insulation.

Of course, thermal storage does have losses over time, but we're talking about hours here, not days. Hours is enough for wind to penetrate to very high levels without too much trouble. Backup (eg biogas) could be used for deep 'storage' like heat waves and such times of elongated low wind production.

Low capital costs, low material input, and the use of commodity materials (almost no rare earths except for some alloys maybe) are things that make CAES concepts promising for large amounts of energy storage and as such worthwhile for further investigation. It's really a pity that there's relatively little R&D on CAES.
 

There are other concepts mentioned as well, such as a traditional diabatic CAES system where the heat from the recuperator comes from CHP - nuclear and geothermal look particularly suitable.
 

I don't think that warm water would help much.  To avoid ice formation in expanding humid air, you'd have to reheat the air every time it got close to freezing.  If you can heat the air to a maximum of 100°C, this will require a lot of reheats (and a similar number of intercoolers on the compression end).  The result would be lots of cost and flow losses, two things you don't want.  The whole point of my concept of evaporating water into the compressed fluid stream is that it would eliminate intercoolers, and the expansion would only require separation of liquid water when the quality (vapor/gas fraction) got down to 90% or so.

We are talking a lot more than a few hours of storage.  The Ridge study assumes 50 hours.  Seasonal energy storage would require much more than this; at least a week, perhaps more.  If regeneration is going to be usable over such timescales, it has to lose heat fairly slowly.  This either means good insulation or large size.  Systems designed for long-term storage of gigawatt-scale energy flows are going to be big regardless, so perhaps nothing special needs to be done here.

Nuclear heat would be good for reheat on a system with non-regenerating intercoolers, and probably good for extending air reserves and expanding total energy output even in a system with good regeneration.  However, I don't see anything like this getting into the testing phase for a minimum of 10 years, and not into commercial use for at least another 10.
 

Sorry for the late reply, just got back from a trip to Hong Kong.

If you can heat the air to a maximum of 100°C, this will require a lot of reheats (and a similar number of intercoolers on the compression end). The result would be lots of cost and flow losses, two things you don't want.

Could you give some figures? Can't be more than a couple hundred $ per kW for a large system. Parasitics may be an issue - what percentage extra losses are we talking about? This would have to be compared to the costs and system losses of other CAES proposals.

We are talking a lot more than a few hours of storage.

Hmm. From a system perspective, long term storage is better accomodated by biofuel/gas backup. Diurnal storage alone would allow large amounts of wind and other intermittent sources on the grid with very little issues. The longer timescale backup would only be used to deal with longer irregularities like large scale heat waves/depressions. Even natural gas or stored coal syngas could be used for now; the biggest gains are from getting the diurnal storage displacing fossil fuels that would otherwise be used for this purpose.

It would get disproportionally less economical to use AACAES for long lasting backup. Higher insulation costs etc combined with lower turnover make such systems relatively costly per kWh delivered. At this point in time it's not what is needed most and the oppertunity costs are high.

You're probably right about the nuclear part. Takes a long time to develop. Still it's something interesting to keep in mind. Cogeneration and trigeneration get way too little attention while the gains are potentially huge.

Thanks for the study. I'll read it when the jetlag ebbs away :)
 

Another crucial aspect to mention with regards to suggesting lower temperature heat storage is the fact that high temperature compressors are inefficient. The hope is that a new higher temperature compressor will be developed that can operate very efficiently at several hundred degrees Celcius. Physics make that difficult, but there may be engineering fixes that could skirt the problem.

Any thoughts? Or maybe some poetry? :)
 

Regarding your main question about residual emissions, I recall a study by which concluded that diabatic CAES (natgas) is uncompetitive under higher carbon taxes. Or very high natgas prices, since they are commensurate. That implies fuel use (and thus CO2 emissions) of traditional CAES are significant. The study did found that transmission (independent wind sites) goes a long way with regards to wind penetration.
 

As for dealing with CO2, mineral sequestration seems promising. Olivine (magnesium silicate) for example, is very abundant and can be mined, crushed to a fine powder, and spread over oceans and farming lands. Projected costs are very low, the process is safe (has been going on for billions of years), less energy intensive than CCS, and has the extra benefit of reducing ocean acidification directly. There's a dutch professor, R.D. Schuiling, who's working on this concept. Very promising stuff.
 

Now it's my turn to apologize for the delays.  I'm swamped with off-line projects and with limited connectivity at the moment.

"From a system perspective, long term storage is better accomodated by biofuel/gas backup."

In a regime of supply constraints, the most desirable system is the one which makes most efficient use of scarce items.  This is where CAES shines.  Wind power may not be geographically distributed where we'd like it, but it is very abundant.  CAES allows fuel, such as natural gas (but possibly biofuels), to produce electricity at an effective efficiency of 80% even in a system without regeneration.  With regeneration, the effective efficiency could be well over unity.

"It would get disproportionally less economical to use AACAES for long lasting backup. Higher insulation costs etc combined with lower turnover make such systems relatively costly per kWh delivered."

The CAES system is divided into the compression/expansion system and the storage.  These two parts are largely independent; the compression/expansion system is sized based on the peak power, and the air storage is sized based on the amount of energy to be held.

A natural air reservoir can be extremely cheap.  If e.g. a a solution-mined salt cavern or fractured sandstone stratum is used as the storage, the incremental expense of storage is very low.  A regenerator made of fractured rock in a mined-out chamber would be insulated by the vast volumes of rock around it, and need no extra insulation.

"high temperature compressors are inefficient."

This is contrary to my understanding.  The isentropic efficiency of modern compressors is around 80%; the inefficiency of the CAES system comes from discarding compression heat on the way to storage.  If a substantial amount of this heat can be saved and restored using regenerators, the system efficiency rises.

(You might note that the 80% value is roughly the same as the gas-to-electric efficiency of the CAES system cited above.  I don't think this is a coincidence.)

As an example, let me do a quickie analysis of a hypothetical CAES system with 80% efficient turbomachinery, 90% pressure recovery in the storage and 75% heat recovery in the regenerator.  It uses air (γ=1.4) and operates at a compressor pressure ratio of 50:1.

The isentropic temperature ratio is 50^[(γ-1)/γ] or 3.06.  With the energy of the losses added as heat, the temperature ratio rises to 3.57 (if the inlet is 288 K or 59 F, the outlet is 1029 K or 1392 F).  (The temperature would actually be lower because γ decreases at such high temperatures, and the temperature rise with it.)  The energy input is proportional to the temperature increase, which is 741 K.

The air storage system recovers 90% of the pressure and 75% of the temperature rise, so the air withdrawn is at 45 atmospheres and 843.6 K (1059 F).  If the air is heated to 1300 K (an increase of 456.4 K), the fuel energy input is equal to 61.6% of the compressor energy input.  If the heated air is expanded back to atmospheric pressure at 80% efficiency, the temperature drop is 689.5 K.  The recovered energy is 93% of the electric input and a whopping 151% of the fuel energy input (net efficiency is 57.6%).

The turbine outlet temperature would be 610.5 K, so there would be some potential for a low-temperature steam cycle to recover more energy (or perhaps for industrial process heat).  If such a cycle can capture 15% of the delta-T between 610.5 K and 288 K, the output increases the net efficiency to 61.6% and the fuel efficiency to 161.6%.

Contrast this to more conventional energy cycles.  The most efficient stand-alone cycle using biomass (a combination of pyrolysis to charcoal and gas followed by SOFC/DCFC to convert to electricity) has potential efficiency around 70%, and its components are not yet available with the sort of industrial reliability we need.  A system using long-term CAES with regeneration can be made with off-the-shelf components and may get more than twice as much energy out of a BTU of biofuel.

"I recall a study by which concluded that diabatic CAES (natgas) is uncompetitive under higher carbon taxes."

Uncompetitive with what?  Even if running on fossil fuel, a non-regenerative CAES can deliver more output kWh per unit of carbon than a stand-alone CCGT plant.  A regenerative CAES can deliver 2.5 times as much energy per unit carbon, and would be highly competitive as a zero-carbon system operating on biofuel.  If the biofuel system captured and stored some of the input carbon, it could be carbon-negative.

"As for dealing with CO2, mineral sequestration seems promising. Olivine (magnesium silicate) for example, is very abundant and can be mined, crushed to a fine powder, and spread over oceans and farming lands."

This effort consumes energy.  Does it capture as much carbon as the energy expenditure would release?  It would probably be better to make negative-carbon energy systems instead.

BTW:  I did a quick calculation on the amount of carbon which CAES systems would take in along with their air.  Even if the USA was heavily dependent on CAES and captured 100% of the CO2 coming in through the compressors, the systems would only remove on the order of a million tons per year.  Grabbing the carbon from biomass would be on the order of 100 times as effective.
 

This comment has been removed by the author.
 

This comment has been removed by the author.
 

Good to see you back EP.

CAES allows fuel, such as natural gas (but possibly biofuels), to produce electricity at an effective efficiency of 80% even in a system without regeneration. With regeneration, the effective efficiency could be well over unity.

True, but it's still a large amount of natural gas. At best it's a 2x-3x reduction over CCGTs in operation today. Large scale implementation will still require vast amounts of natural gas.

Because as you say the marginal storage costs in a traditional CAES system are low, it could be a good idea to have an AACAES (which has higher marginal storage costs due to the extra cost of thermal storage) to deal with diurnal variations and and a diabatic CAES system as deeper backup (for heat waves and such).

The CAES system is divided into the compression/expansion system and the storage. These two parts are largely independent; the compression/expansion system is sized based on the peak power, and the air storage is sized based on the amount of energy to be held.

I was talking about AACAES with a seperate thermal store, but you must be talking about your steam integration idea. But if either of these is not properly insulated, it will lose too much to be viable as deep backup. The ammonia thermochemical storage does not suffer heat losses over time (it's not latent heat) so could be a solution. But, it's in the prototype stage right now and so far it appears very expensive.

"high temperature compressors are inefficient."

This is contrary to my understanding.

In theory, higher compressor operating temperature should be more efficient. The problem is that there's no good proven technology for high efficiency high temperature compressors (say 600 plus degrees C), although they are under development. Problems with materials (ceramics are being tried) such as creeping and other durability issues are some of the problems. Heat management was also a problem. To my knowledge, very high temperature compressors that are also very efficient were never really necessary so there was little incentive to solve the problems. But maybe 200-300 degrees C or less shouldn't be too big a problem. So for subcritical water storage it should be doable to create a cheap and efficient system.

I recall a study by which concluded that diabatic CAES (natgas) is uncompetitive under higher carbon taxes."

Uncompetitive with what?

With HVDC long distance transmission. To take advantage of decorrelated wind regimes. There's a strong trade-off between storage and such long distance transmission, although it's not a 100% trade-off.

This effort consumes energy. Does it capture as much carbon as the energy expenditure would release? It would probably be better to make negative-carbon energy systems instead.

We should do both. But olivine has real potential. A ton of olivine captures a bit more than a ton of CO2. I think that sequestering 30 GT/year this way would require a few (2-4) percent of world oil production, based on Ian McClatchies estimate. A large amount of oil but not impossible.

Ian McClatchie thinks it would require roughly a gallon of diesel per ton of olivine on the basis of 100 mile transport (probably true since large olivine deposits exist near the oceans). This doesn't include crushing and mining but then again a ton of olivine sequesters more than a ton of CO2. The operations could be powered by a dedicated nuclear powerplant. Electrified mining equipment and rail. There would be a relatively small fuel requirement for ships because they would not have to travel far.

BTW: I did a quick calculation on the amount of carbon which CAES systems would take in along with their air. Even if the USA was heavily dependent on CAES and captured 100% of the CO2 coming in through the compressors, the systems would only remove on the order of a million tons per year. Grabbing the carbon from biomass would be on the order of 100 times as effective.

That's interesting. But what bothers me is that the traditional CAES system would require very large amounts of natural gas. Biogas would work but that's also a lot of biogas.

Current global CO2e emissions amount to more than 27 GT/year. Capturing 100 MT/year via biomass is only 1/270 of total emissions sequestered. Worldwide potential would be bigger though. Perhaps a few percent of today's annual emissions. Great, but we need more than that.

It took millions of years for fossil fuels such as coal, natural gas and oil to form, and we are burning it on the scale of centuries. There are caveats here, such as a lack of suitable conditions for carbonization and re-oxidation of deposits back into the atmosphere. But even then, this is very indicative for me - plants are not very swift in sequestering carbon.

Olivine has big potential. This type of process has been happening for billions of years already, at a rate of more than a million tons per day (on the top of my head - don't know the exact figure). It's called erosion. Reaction of minerals with acidic CO2.

Two cubic miles of olivine mined, crushed, and deposited into the rivers, seas and oceans would sequester more than 30 GT/year. A non trivial amount I'd admit, but doable (total global human mining activities far exceed two cubic miles per annum).

Cheers, Cyril
 

" Olivine ... can be mined, crushed to a fine powder, and spread over oceans and farming lands."

This effort consumes energy. Does it capture as much carbon as the energy expenditure would release?

Way more. About ten times more. I looked at pulverization energy (50 electrical kWh per tonne) and the energy to fling the powder 5 km into the air and it came to about one-eighth of the electricity formerly yielded by coal-fired power plant in putting up the CO2 that this takes down.

I was figuring only on Mg2SiO4 going to two MgCO3s, solid, but others closer to the action say the carbon atoms pulled down per mole of forsterite (olivine actually is a varying mix of forsterite (Mg2SiO4) and fayalite (Fe2SiO4)) is four because CO3^(2-) in the ocean pulls down most of another carbon to form bicarbonate.

More at RealClimate.