Comment by usrusr
3 days ago
I wonder how much Google is factoring in the implicit cooling cycle? Because any pressurized gas energy storage is either including some advanced heat storage or is just venting the heat created during compression (the ancient Huntorf facility in Germany is infamous for that, super wasteful)
Usually you want to keep the heat and put it back into the compression medium during decompression and hope that losses from the heat storage aren't too big, but when you have a cooling use case nearby, you can use that low intensity heat to compensate heat storage losses, or even overcompensate. When you consider how much of the power input of a datacenter is typically used for cooling, compressed gas storage could be useful even if there was zero electric recovery (just time-shifting the power consumption for cooling to a time with better energy availability)
> Because any pressurized gas energy storage is either including some advanced heat storage or is just venting the heat created during compression
https://energydome.com/co2-battery/ diagram has water as the heat storage. Tanks of water get efficient at energy storage due to square-cube scaling.
I'm sort of thinking out loud here but could you have two batteries running simultaneously but on opposite cycles, so while one is cooling the other is heating? Obviously it wouldn't be 100% efficient but it might reduce some wasted energy.
The heat and cold are created by the compressing or decompressing the CO2 (our any other gas). If one battery is heating while the other needs heat that would imply that one is charging while the other discharges, which is rarely useful in normal operation
Yeah you're right I didn't think that through, why would you charge a battery while discharging another, just use the energy directly.
Why would you be charging one battery while discharging another? That would just be wasting energy.
If Google is colocating these with data centers, even low-grade heat that would otherwise be a loss could still be useful, or at least reduce how much active cooling the DC needs
Isn't this effectively neutral over time? Heat generated during compression, lost during decompression, so basically using the air as a heat storage medium?
I think what he's saying is you can boost efficiency if you compress a cooler gas. So if you could capture the "cold" that you get from discharging the device, and use it to pre-cool the air for the next cycle (or use it for the data centers cooling system) , it would be much more efficient.
Cooling is rarely done in any other way than compressing a gas, allowing the heat to dissipate and then allowing it to decompress again. You don't want to compress a gas to cool another gas about to be compressed. What reasonably advanced compressed gas storage systems do is capture and store the heat that gets created during compression and feed it back during decompression. This gives the same efficiency difference as compressing some magically pre-cooled gas would do, only on the discharge side.
So far so good, just the old thermodynamics. It gets interesting when you have a cooling use case anyways: then you can skip on some of the decompression recovery and use the "cold" from decompression directly, to cool down something that needs cooling, without going the extra way of converting back to electricity and then sending the electricity recovered into a compressor setup for "creating cold". Bonus points if you also have a use case for the heat you did not use in reconversion to electricity, but chances are between losses during storage and heating some the gas back some amount beyond neutral you won't have much spare heat anyways.
Yes. A large radiator would handle both. I assume they just store the heat because hot water will be a lot more efficient at reheating the co2 than night time air and a pool with an insulated cover is not hard to construct.