You are pedantically correct sir, but there is a larger context. So in that larger context I was working on this plan when working for a company that had a data center near the Columbia River and 'industry leading' PUE numbers.

Generally when the wind was blowing (and it blows a lot along the Columbia River Gorge) there is no need for additional cooling. When there is no air movement and the outside temperature is high, the demand for cooling increases requiring a more 'active' engagement (sorry but I can't be more specific than that).

So realistically I was looking to time shift the excess cooling capacity that was available during windy times to the times when there was little wind and excess heat. Compressing air has some advantages in that the amount of energy you can get out vs the amount you put in is comparatively efficient to battery power, and using / exploiting the cooling effect of the expanding air, and the kinetic energy of its expansion to generate electricity, was actually more efficient than other methods. Maintenance burdens were also lower and municipal permitting was made easier by the lack of 'dangerous' chemicals or catalysts.

These advantages are perhaps uniquely suited to data centers. Of course I didn't get to actually build it because MDI wouldn't participate. That was too bad but certainly well within their rights.

Essentially what you are saying is that your want to (also) store cold rather than store just energy.

I'm not convinced it was worth it. You anyway are collecting energy from the wind, and instead of storing it as energy you stored (part of) it as (potential) cold.

But in order to do that you threw some of the energy away - I suspect that if you stored all the energy directly, and then created the cold on the spot as needed you would have higher efficiency.

If I did want to store cold I would not do it using compressed air, I would use a water tank - cool the water to ambient (no active cooling, just some pumping) then dump heat into the water when necessary.

Perhaps when I get a chance to build my own data center for Blekko I'll be able to actually run the experiment [1]. Then I'll have a better answer. But would like to share a bit my reasoning.

You are correct that my concept 'threw away' the heat generated by compressing the air. However, since any energy stored at all was collected from a previously unharnessed source (the wind blowing past the data center) it might be more accurate to say that it would not collect all that it 'could'. But here is a very important distinction, the heat from air compression is 'low grade' heat, which is to say the delta between it and ambient is small enough that harvesting it in meaningful amounts is quite difficult. You also need to consider that this thing is operating next to a data center which is running an evaporative cooler 24/7 to pull excess heat from inside the box and dump it into the air outside the box. Basically spending energy to move the heat outside.

So from this side of the screen, to accept you point that not capturing the heat generated from compression was 'throwing it away' I would need to have some credible way of utilizing the captured heat.

Looking at it from the overall energy exchange picture, you've got the kinetic energy of the wind, with is 1/2mv^2 where m is the mass of the wind moving pass the windmill. Some percentage of that you convert to mechanical energy which runs the air pumps and some of that becomes heat in the bearings and linkages. The mechanical energy then compresses the air which by the ideal gas law goes up in temperature proportional to the change in volume. Just sitting there, the tanks holding the air cool off, losing this heat to the surrounding atmosphere. We could insulate the tanks to keep it inside (and pressures up) but it turns out that we'd like to harvest heat energy out of the data center later so we let this heat leave by convection. The tanks then go down in pressure thanks to that same ideal gas law.

So when we use this air, we decompress it inside our data center. This allows heat in the data center to be absorbed by the expanding air, which saves us the energy we would have been using to run the evaporative coolers to pull it out and increases this energy available in the air as it increases that static pressure differential between ambient and the decompressed air. So we 'get it back' as it were, in a positive way. Running it through the engines allows us to then harness the pressure differential, and if it comes out of the engines still 'colder' than the data center air we can inject it right into the regular atmosphere of the data center to offset warm air that is already there.

[1] Facebook also has a data center up there in the region and if someone there wants to pick up the ball and run with it that would be pretty cool to. You could write it up for the open compute stuff. Of course you still have to figure out how to get an MDI engine. If Tata ever shipped I considered buying cars, throwing away the body and just pulling out the engines. That would work but requires Tata to actually have a product you can buy.

This is an awesome idea by the way. I work in the energy industry and haven't heard about anything like this.

Just curious, it seems like with prior art, MDI would not have exclusive rights to the motor, right? Also their Wikipedia page says that they are currently using a design that was patented in 1990. Did you abandon this idea because of patent issues, or was it merely the lack of access to a proven/tested motor that stopped you?

It was getting reasonably efficient compressed air motors of decent size. The MDI docs claim they do about 50hp (or about 37kW). You'd probably want five or ten 'units' at that level (assuming the data center is 10 or 20MW like the ones Apple, Facebook, and Amazon have built).

You can of course run an existing steam engine on air pressure (I've done that with models) but they are optimized for a higher static pressure than the MDI engines target.

Not for all possible thermodynamic cycles -- there is no fundamental thermodynamic irreversibility due to compression or expansion.