A company is building a giant compressed-air battery in the Australian outback
wired.com
wired.com
What a compressed gas represents is not stored energy, but stored (negative) entropy. It is a resource that allows low grade heat to be converted to work at high efficiency. This is what's to happen in this facility: the heat of compression is separated out and stored, then used to reheat the compressed air at discharge time. The energy is actually being stored in that thermal store.
But there are other ways to do this that don't involve compressed air storage. Instead, after the heat of compression is removed and stored the compressed air could be reexpanded, recovering some of the work. This would leave the gas much colder than when it started. This cold could be stored (heating the gas back to its initial temperature) and the gas sent around again. To discharge, the temperature difference between the hot and cold stores could be exploited.
This is called "pumped thermal storage". I believe Google/Alphabet has/had a group looking at this (called Malta). It has no geographical limitations.
(*) Highly compressed air will store some energy because the molecules become so crowded some energy is stored in intermolecular repulsion, but that should be a small effect in this system.
The effect the prior poster was discussing is most apparent at large scales when someone is doing actual efficiency analysis - compressed air is not a very efficient way of storing or transporting energy in most contexts. No one would use compressed air for a power grid, because it would quickly be apparent how terrible it is.
It is however a great way to transfer and use energy in many industrial contexts, as turbines and pistons using high pressure compressed air can be very compact while also being very powerful, and have built in cooling.
At the level of typical household use, the efficiency effects aren’t notable. No one is going to care or even notice if it cost 1/2 cent of electricity (at the compressor) to use that air impact wrench vs 1/10th of a cent to use an electric one. Especially when the electric one is heavier and has less torque.
do you know how they do it?
This sort of design (a counterflow heat exchanger, basically) keeps the delta-T at each point low, so the system has low entropy production.
A similar heat exchanger could be designed with a liquid thermal storage medium, where the air and the liquid are flowed past each other in opposite directions (separated by solid walls, most likely). The hot storage tank would be insulated. An advantage of this design is the pressure of the liquid storage tanks needn't be high, unlike the vessel containing the pebbles in the earlier described system.
All these may need to dump some waste heat to the environment as a consequence of inefficiencies in the system.
Compressing air absolutely does store energy. It takes work to compress it, and work will be done when it is allowed to expand.
Yes, one way of measuring the stored energy is by the resulting increased temperature. Compressing air necessarily raises its temperature. And then if you want to you can transfer that heat elsewhere, go ahead. That's how air conditioners and refrigerators work, after all.
But in the basic case, it seems entirely accurate to say that compressing air does actually store energy. Just like raising a heavy object against gravity stores energy.
Obviously if you let the compressed air cool, then you're letting the energy dissipate. So if you want to preserve all the energy, don't do that.
The oh-so-clever trick is to transfer heat elsewhere while you're compressing the gas, so you actually reduce back-loading on the piston instead of having it 'fight' the temperature rise. This can be accomplished via water spray, or by compressing a gas bubble that's surrounded by water (eg in the trompe).
By continuously removing heat as you compress the gas, it effectively acts like a train of compressors and intercoolers with an infinite number of stages, ie true isothermal compression. No magic, just physics.
This is unlike traditional CAES, where the air is reheated by burning a fuel. There, fuel use is minimized by isothermal compression.
It seems like either this is describing an additional energy storage/extraction method that’s possible with these systems. Either to improve the efficiency or something that’s actually far more efficient than utilizing kinetic energy compressed air can create.
It’s kind of like saying the water in a hydropower plant doesn’t store energy. The turbines aren’t extracting energy from the water itself. They’re extracting energy from gravity acting on water. It doesn’t matter if the water is nearly freezing or nearly boiling, it has roughly the same amount of hydropower available.
This is not the case in a CAES system. The internal energy of air, which is the kinetic energy of its molecules, is independent of pressure. It's only a function of temperature. The internal energy of the air is unchanged by a process that increases its pressure without increasing temperature. So, no energy has been stored in the air.
What has been done is reduction of the entropy of the air. So, it is now possible to convert some of that internal energy (which was always there) to work by expanding the air. The air would end up even colder than the initial air.
https://www.youtube.com/watch?v=50fJ8Av_g7Q
https://www.youtube.com/watch?v=uvf0lD5xzH0
https://news.ycombinator.com/item?id=27066295
Sad to see people are misinterpreting the cleverness of isothermal compression, and thinking it's "just" thermal storage.
How the conditions are reached makes a big difference in efficiency.
I'm baffled at why you focus on minor implementation details like internal energy at a particular snapshot in time, as opposed to highly salient metrics like overall round-trip system efficiency.
What isothermal CAES (with heat storage) would require is a larger storage volume per unit of energy storage capacity, since less heat is generated during the ideal compression. That heat generation is not inefficiency, it's something you want!
Only if the heat is discarded and you reheat with fuel do you benefit from isothermal compression.
I have an injector pump that sits at the top of the borehole and has a pipe that pushes water down to pump water up through a second pipe, but it's more common to have a submersible pump down at the bottom of the bore.
What is at the bottom of the hole is a "injector" which is basically a U shaped pipe and a small jet to push water back up. If the well water is only 25 feet below the ground a pump at the top along works. This jet system gets down to 60 feet. The pump down the hole gets to 600 feet (check the pump specs - many are rated to only 250). After that you need oil well type pumps where the motor is at the top of the hole but the pump is lowered down.
https://www.vvsbutiken.nu/product.html/ejektorer-grundfos?ca...
Two normal plastic water pipes descend down to the bottom of the borehole. At the bottom is a brass u-shaped fitting with a inlet with a non-return valve.
You might think that you'd need a large amount of water to make the energy release work, but I think it works like this. The force of the water on the air/water interface is dependent not on the reservoir volume, but on the weight of the water in the column (which depends only on the height of the shaft).
By digging a very deep shaft, you can have a very large force of water on the interface, and moving that interface an equal distance hence releases more energy than it would with a shorter shaft.
This way, you can store an arbitrary (up to your ability to compress air to a sufficient pressure, dig a deep shaft, and keep everything from blowing up) amount of energy.
I think if you have a 1 square meter cross section of shaft, and the shaft is 1 kilometer deep, then at the bottom the force of the water above is the weight of 1 square meter * 1 kilometer, or 1000 cubic meters of water, or 1 million kg.
The force then is 9.8×10^6 newtons.
Pressure is force/area, or 9.8×10^6 newtons/square meter here (since we have a unit area).
There's a formula for the energy in compressed air, it's... involved. I downloaded an excel file from here: https://ehs.berkeley.edu/publications/calculating-stored-ene...
It says under this pressure, a 1000 cubic meter tank of air at this pressure stores ~5000 KWh.
The one planned in California is supposed to store 4000 MWh, so I guess they have a tank that is ~a million cubic meters.
A water tank of the same size would store ~2700 MWh of energy (e.g., pumping that much water up a 1KM shaft requires that much energy), so it does seem to be more efficient.
It may also be that hot air turbines are cheaper and easier to maintain than hydropower turbines, but I'm not certain.
+ they are storing the heat extracted from compressing the air.
Compress to store [potential] energy, take the heat generated during compression and use it, decompress at time of need and let the [hot?] atmosphere supply energy to the decompressing gas.
I wouldn’t call it a closed loop, but maybe a lack of deficit.
Just wanted to highlight this, since it's the key insight which caused this to "click" for me.
The difference between a 1km shaft and a 1km deep reservoir, when both are used for pumped hydro, is the amount of energy stored. Running a turbine at the bottom of the shaft (to where?!) would drain very quickly. But using air, which has very different properties, enables you to use the water as a "piston" to keep the air at a certain constant pressure underground. And you can store a larger volume of (more compressible) air in a space which is easier to access.
You could probably move the compressed air in a GPE gravitational potential energy storage system, or haul it up on a winch and set it on a shelf; but would the lateral vectors due to thrust from predictable leakage change the safety liabilities?
Air with extra CO2 is less of an accelerant, but at what concentration of CO2 does the facility need air tanks for hazard procedures?
FWIU, you can also get energy from a CO2 gradient: "Proof-of-concept nanogenerator turns CO₂ into sustainable power" (2024) https://news.ycombinator.com/item?id=40079784
And, CO2 + Lignin => Better than plastic; "CO2 and Lignin-Based Sustainable Polymers with Closed-Loop Chemical Recycling" (2024) https://news.ycombinator.com/item?id=40079540
From "Oxxcu, converting CO₂ into fuels, chemicals and plastics" https://news.ycombinator.com/item?id=39111825 :
> "Solar energy can now be stored for up to 18 years [with the Szilard-Chalmers MOST process], say scientists"
> [...] Though, you could do CAES with captured CO2 and it would be less of an accelerant than standard compressed air. How many CO2 fire extinguishers can be filled and shipped off-site per day?
> Can CO2 can be made into QA'd [graphene] air filters for [onsite] [flue] capture?
"Geothermal may beat batteries for energy storage" (2022) https://news.ycombinator.com/item?id=33288586 :
> FWIU China has the first 100MW CAES plant; and it uses some external energy - not a trompe or geothermal (?) - to help compress air on a FWIU currently ~one-floor facility.
Also, 1000m3 in air is just a cube of 10m. It seems like a good illustration of the convenience of air vs water.
Edit: Also the fact, that if the pump was at the very bottom and it failed, you'd need to have an alternative way to clear the cavern of water in order to go down and fix the pump.
Put the pump on a chain and hoist it up with a crane, like they do with pumps in lift stations.
Not only can you store the energy required to lift the water to the surface, but after all the water is lifted you can keep compressing the air in the closed reservoir to store even more energy.
If you only used water you could of course build an open reservoir with more storage capacity instead. But maybe these will be built in areas where that’s not feasible. It’d take much more space on the surface, and you need to deal with evaporation.
Constant volume CAES would store energy without any water lift involved, and with water mediated constant pressure CAES the lifted water is added to the amount of energy contained in the "air spring". But that's balanced at equal force in the force x surface area x underground reservoir level system. My bird's eye view understanding suggests that it would come down to 2x the amount of energy stored in either (minus losses, of course) due to the balance. Is that the maximum energy a water-column mediated constant pressure A-CAES could hold, per water displacement volume x head height, or am I missing something? That 2x would surely be an improvement over conventional mineshaft pumped hydro, but it would also define a somewhat sobering limit to the amount of theoretical best case capacity.
Minor but perhaps very relevant detail: afaik, or rather as far as I don't know, hydrostore aims at a shaft depth equivalent to a pressure either right below of where pure CO2 would liquefy, or right above where that happens (no idea about the boiling points of N2, O2 and all those other main components of ambient air). I think the target depth is not quite that deep, as in carefully avoiding the "transliquid" range (as in transonic).
1. It takes up less space on the surface than most PHES. This... is almost always a marginal benefit.
2. It doesn't require building a reservoir or dam. These are very well regulated in Australia and elsewhere, and the downsides are known so they are quite slow to get approval.
3. It's a bit quicker at 2.5 - 3.5 years as opposed to 3-7 years.[1] This is a bigger advantage than it looks if you have some tricky renewable energy targets to hit by 2030 (see our 42% emissions reductions target as well as an 82% renewable energy target)
I can't see this gaining traction outside of a few locations in Australia, at least. I wouldn't be surprised if A-CAES is only briefly viable as a result of subsidies and cheap government financing.
[0]: https://re100.anu.edu.au/#share=g-fa5a20c9c63f6ed6343a7e7573...
[1]: https://www.csiro.au/-/media/Do-Business/Files/Futures/23-00...
How so? Yes, if you dammed up most valleys you could build a lot more capacity than humanity currently has, but that's because we don't really have that much. A factor of two or so might be well within range. But the total amount reasonably buildable simply isn't enough, except for some very local scopes where demand is low in both energy and in other use for the landscape that would be taken over by reservoirs. And that's before you start considering the geological realities required for actually building a dam, you don't just need the geometric shape of a valley, you need the bedrock to brace the dam against and the impermeability of the ground required to not have leakage wash out pathways for ever increasing leakage. Viable sites for pumped hydro are extremely rare and quite a few have actually been given up decades after building the dam, because the geology wasn't quite as cooperative as hoped.
The promise of deep site water head CAES is that you can just throw money at the problem (excavation) and get as much capacity as you want to buy. The price per capacity is higher than that of a low hanging fruit pumped hydro site, but many of those buildable have already been built.
For example: https://www.news.com.au/technology/environment/abandoned-min...
(I agree that this is different from traditional hydro locations, and so there are many good spots still available)
50 year old example: https://en.wikipedia.org/wiki/Ludington_Pumped_Storage_Power...
1. Dams require water. Australia has suffered water supply challenges as long as I've been alive.
2. Dams can be environmental disasters. Both building, maintaining and one day destroying these has a lot of challenges and expenses that we're not great at measuring.
1) You're not losing thermal energy from the main body of air (assuming their thermal recovery at the compressor works well enough) because the change in air pressure only occurs at the compressor.
2) You get constant pressure and constant power output for the entire volume of your compressed air storage, instead of both dropping rapidly as air is released. This gives you far better bang for buck per unit volume of pressurized air.
I haven't done the calcs but I'd guess that if they're bothering doing all this extra stuff the energy stored in a cubic meter of compressed air at these pressures is significantly higher than the energy you'd get from just lowering a cubic meter of water down to the reservoir.
(see nephew comment https://news.ycombinator.com/item?id=40273030 )
The question is how much more, whether the symmetry suggested by the balanced state implies 2x energy storage or not.
Another mental model to further separate the two storage media (water lift and air compression) is this: you could charge separately, first raising the water with a pump, with the dry cavity at ambient pressure via an open connection to daylight, then seal the cavity an charge it as a compression vessel in the non-isobaric way. I still can't decide between trusting the "balance implies 2x" intuition or not, but it should be easy enough to calculate. At least the thought experiments show that it can't be 1x the energy stored by lifting water alone.
I consider starting with "let's do compressed air energy storage" as a given, the "do it by pumped hydro" does not come into play because pressure vessels are hard, it comes into play because building compressors and turbines that have good efficiency over a wide range of pressure is hard.
But you don't arrive at isobaric storage looking at mineshaft pumped hydro wondering if it could be improved upon. You start looking at constant volume (A-)CAES, either in high pressure tanks at/near the surface or in salt caverns, and go from there, what of we did not have to deal with a very wide range of operation pressure.
In contrast, there is underground everywhere on Earth.
Pumped hydro done on a large scale needs a reservoir that is at a considerable elevation gain.
example: https://en.wikipedia.org/wiki/Taum_Sauk_Hydroelectric_Power_...
Their California battery will be 4GWh capacity with a $1.5 billion cost, which is $375/kWh. Their Australian one will be 1.6GWh for $415 million USD, working out to be $260/kWh. Both are more expensive than lithium ion, so I wonder what the case is for it.
1. no degradation
2. cheap to expand? - simply expand the cave
That is not cheap. And we have very high pressure here and not only cave and rock, but technic around it. And pushing air in and letting air out again will have degradation of that expensive equipment.
Are you comparing battery cell cost vs battery pack + structures + electronics + lines + land + installation + different continent + N other things I have no idea about?
https://about.bnef.com/blog/global-energy-storage-market-rec...
And how long do lithium batteries last?
A battery UPS under my desk doesn’t really affect my rent or mortgage. Buildings not only need to get built they also need to be maintained.
So both longevity and working towards reduced costs for future plants. I guess someone thinks the long-term costs will end up below Li-ion, and likely lower environmental impact, at least compared to the battery component.
In the end 3 figures really matter total capacity, power output and cost per "generated" kWh on average over lifetime.
The reservoirs will at some point see noticeable sediment buildup, but not at all comparable with surface pumped hydro based on blocking valleys, due to the cyclic nature of the water flow in the hydrostor facility. And occasional cleanout (measured in decades or centuries?) will be trivially cheap compared to construction cost. Very much unlike cleanout cost behind valley dams, which are would-be cleanout costs because that never ever happens as it would utterly dwarf the cost of the original dam. There's been an article linked here a few months ago (can't find it, unfortunately) about how the total capacity of pumped hydro is getting increasingly smaller each year, despite new sites getting built. This is because sedimentation already outpaces the buildup, and it will only get worse the more we build. The volumes accumulating behind a dam are just too big, unless you have zero natural flow from rainwater (and then just getting the working volume of water to the site would be prohibitively expensive, per capacity, even before you factor in evaporation - the cycle capacity per unit of water is just so much lower than what a hydrostor site would achieve with its much larger head plus the energy stored in air compressionand heat)
Another aspect, completely unrelated and I'm not sure hydrostor already makes that part of the design (but it could totally be introduced later, without invalidating any of the excavation investment): some of the energy stored in an A-CAES system is stored as heat. When you do need heat, for a district heating system, for a swimming pool site or whatever, you can decouple some of the heat from the pressure storage. Worst case some of the joules repurposed end up missing in discharge, but it's also possible that they are simply joules not lost to cycle inefficiency. And if you happen to need cooling (datacenter on site?), at the time of discharge you can just keep some of the stored heat untapped, substituting with energy from the warm end of the coolant cycle that you want to freeload on the A-CAES. Compared to other waste heat/cold coupling schemes, at hydrostor pressure levels you would get considerably higher heat/cold deltas to work with. Huge potential, and with the reservoir shaft having very few site requirements, coupling opportunities should be plentiful (as compared to e.g. opportunities that only ever arise in remote valleys)
The article says this thing should last at least fifty years. There's no good reason it couldn't last longer as it shouldn't really degrade over time. If you assume daily charge/discharge cycles, that would be about 18250K cycles over 50 years. Times 4GWh is about 73TWh of energy sold to the grid at, hopefully, some profit. Of course that all depends on demand, utilization, and whether there are any cheaper ways to store energy. It's probably going to end up some percentage of that. But best case that's energy you buy cheap and sell at a higher price. Even a few cents difference starts adding up to billions pretty quickly. And that's before you consider the alternatives (buying energy on the open market from another provider, investing in more energy generation, etc.).
The prices you cite are just the purchase price. And of course lithium batteries don't last forever. So you'd be writing them off at some point. But in fairness, there are some battery chemistries that are getting quite good cycle times. So, the comparison might become a bit more fair over time.
That there is a finite supply of Lithium available on Earth?
It would take a long time to get it up to initial pressure, as there would be a lot of heat to dissipate, but then it differential mode, the gradient would be much better.
Does having an increase in the density of gas present an advantage over having a higher pressure delta by dumping to ambient for a given volume of compressed-gas storage?
Why would I want one tank at "high" pressure, and another tank at "higher" pressure, when I could just have one tank of "higher" pressure to begin with? Or, better: Two tanks of "higher" pressure in even less space than one of "high" and one of "higher" pressure?
(If the answer is "Because turbines work better with higher densities," then: Do they work more-betterer-enough to make up for the size and complexity?)
The first is fundamental: density is not really helpful for this sort of application. The work done in expanding material (including gas) at a given pressure is P dV, and the work done in moving material across a pressure difference is V dP. Notably, mass doesn’t appear at all here, so adding more mass or density doesn’t add energy storage capacity in and of itself.
You can compute this more explicitly, and, for an ideal gas, the useful energy extractable from a tank of gas at pressure P (under ideal, isothermal conditions) is proportional to the change in log P. So it’s actually rather more important to achieve low pressure than high pressure.
On top of this, there’s a practical consideration: the atmosphere is an effectively infinite source of gas at 1 atm. If you are working between high and higher pressure, you need two reservoirs.
All you’re gaining is less temperature change per unit pressure change, but it’s probably the same amount of heat for any practical purpose, so you still need an intercooler of some sort for good efficiency.
If you were able to keep the air hot the whole time the process is almost symmetrical so that's not an issue, the “heat problem” as I call it is how do you store this heat for an extended period of time? At scale, it's much harder to keep than just the pressurized air.
The prototypes I've seen in the past were not storing the heat, but relied on industrial fatal heat (that was lost anyway) but this also has scale problem as you don't have that much available power except near very specific industries (NPP are an option, as are other heavy industries, but the supply is necessarily limited)
The positive aspect of such a system is that the thermal energy you need is not subject to Carnot's law, so the temperature of the heat source doesn't matter unlike most use of thermal energy (and that's why you can use waste thermal energy in the first place) but you still need a way to get that energy.
One advantage this has (I assume) is almost limitless cycle lifespan.
The idea is that this is a hedge: if it turns out that solar/wind become "too cheap to meter"? Then "efficiency" is meaningless and what matters is cost-per-unit-storage and the hope is that compressed-air will be able to store and output more joules-per-dollar than any other storage method (regardless of how many more joules you had to put in first).
But you are right in that this number really isn't all that important in the renewables age: when we are anywhere close to getting to 100% renewables on a point of median supply and median demand, production capacity (conversion capacity from sunshine and air movement) at times when sun coincides with wind will so far outpace demand that any energy sink will do that can still pay a positive rate. We're a long way from fully renewable were I live, and some days I see two thirds of turbines stopped in nice wind. It's really all about capex per W and per Wh. Admittedly, A-CAES isn't necessarily excellent in this right now, but it might scale quite nicely with routine.
Recently there was a gravity storage scheme linked here on hn about lowering mining refuse back into old mines to discharge, and digging it back up to charge, which comes with the curious property that it never really reaches a point of saturated capacity: in theory you could keep digging new tunnels forever when energy surplus keeps coming in.
> When it’s time to discharge energy, the system releases water into the cavern, forcing the air to the surface. The air then mixes with heat that the plant stored when the air was compressing, and this hot, dense air passes through a turbine to make electricity.
By "releases water into the cavern", do they mean simply opening the air valve to let the air (pressured by the water) come back out?
I'm actually pretty excited about this tech - it seems like solar and wind are getting cheap faster than batteries will be able to meet the needs for grid-scale energy storage, so a cheap-but-inefficient energy storage tech is an exciting prospect. Massively overbuild the solar/wind and use these things to defer the overflow.
They probably won't even have a Thunderdome. =(
But there might another way to avoid bubbling which I have missed. Maybe the compressed air is under so much pressure that it becomes denser than the water?
Imagine an cavern shaped like the great pyramid of Giza, full of water. They dig two pipes, an air entry pipe to the point at the top, and a water exit pipe to one of the corners at the base. Air pumped in at the top pipe will almost evacuate the cavern, before the water level drops to the point where air could get to the exit pipe.
What this means is that depending on the shape of the cavern, they may not be able to utilise its whole volume. In the worst case, if its roof were entirely flat, they would not be able to use any. They can use the volume of a section, bounded by horizontal planes, from the air entry pipe either down as far as the exit pipe, or (if the roof dips down between the two) down as far as the lowest point on the highest possible path between the two.
That's assuming totally vertical pipes. I know they can curve them a bit, but I don't know if they can curve them enough to enter a cavern from below. If they could, then they can utilise the entire volume, as long as they can identify the lowest and highest points
Nothing personal to the commenter I'm replying to, but I love watching HN re-engineer long existing basic systems.
https://en.m.wikipedia.org/wiki/Flywheel_storage_power_syste...
I also remember reading about a system that moved dirt/rock up a mountain on a train. Can’t find a link, but that also seems capital intensive and requires different geology.
There is also pumped hydro storage that works via gravity. That’s been around a while. My dad worked as an engineer on one in the 80s [1].
[0] https://www.energyvault.com/ [1] https://en.m.wikipedia.org/wiki/Helms_Pumped_Storage_Plant
. pipe
| | | |
|---|-|---water level --|
| |_| pipe end below |
|_______________________|