I am working on a compressed air energy storage system. The size and weight of the system are heavily constrained; nothing should (ideally) exceed a few pounds. For this reason, I would like to store the energy (compress the gas) and extract the energy (make the gas do work) with the same mechanism in a rotary fashion.

Essentially what I need is a rotary air compressor that, when air is forced through in the opposite direction, doubles as a pneumatic motor. I'm working with fairly high pressures (I'm estimating a few hundred psi) but low volume. In my search, I have found a plethora of compact rotary air compressors and rotary pneumatic motors, but there is hardly comment on what systems would work as both.

To me, it seems very intuitive that an air compressor could have these properties, but I don't want to jump to any conclusions. I have looked at several compressors, and the most applicable to my situation seem to be:

  • Centrifugal compressor
  • Axial flow compressor
  • Rotary screw compressor
  • Rotary vane compressor

The centrifugal compressor is ideal, but it seems the least likely in my eyes to be reversible, at least with any efficiency. I also looked at pneumatic motors, of which there were fewer available. Most applicable seemed to be the:

  • Rotary vane motor

Other systems, such as the pietro motor, were obviously not applicable in my light weight, compact application. The correlation between the rotary vane compressor and the rotary vane motor is promising, but I would like to know about any options I have.

What rotary gas compression systems can double as motors powered by the gas they compress?

EDIT The answer most likely lies in the similarity between a radial (centripetal) turbine and a centrifugal compressor.

  • $\begingroup$ And a piston-valve compressor? $\endgroup$ Commented May 29, 2015 at 8:08
  • $\begingroup$ @ratchetfreak Good thinking. The piston-valve compressor would probably suit this need nicely. Alas, the piston-valve compressor is reciprocating, which means more vibration and (not necessarily, but I imagine practically) weight than my constraints will really allow. I'm looking for the smoothness of a rotary compressor. $\endgroup$
    – MikeJava
    Commented May 29, 2015 at 15:06
  • $\begingroup$ Low volume, high head applications generally call for an impulse turbine for rotary turbines. Since that seems to be very difficult to turn into a compressor as well, I would go for @ratchetfreak 's option. If I were you, I would focus my research on cars running on compressed air, since I think they may already have developed the exact system you're describing, and if I'm not mistaken, they indeed use a piston-valve system. $\endgroup$
    – Sanchises
    Commented Jun 6, 2015 at 20:09
  • $\begingroup$ @sanchises I don't see why high pressure, low volume applications couldn't be met by any variety of rotary air motor. Why do you say an impulse turbine is necessary? I must be missing some facet of the situation. Also, I would definitely not be surprised if many emerging automotive technologies adopted a piston valve system, it makes the most sense for them. But this question is really about a rotary system, the advantages of which are many. I'll edit the question to make this more clear. $\endgroup$
    – MikeJava
    Commented Jun 6, 2015 at 20:49
  • $\begingroup$ @MikeJava I'm not saying it's not possible - I'm just saying that the most efficient way of using a rotary turbine for a low-volume application is a.f.a.i.k. an impulse turbine, of which there are many options (often just related to how exactly the turbine vanes are designed). I can't find my book on that though, so I guess somebody else will have to come up with the definite answer. $\endgroup$
    – Sanchises
    Commented Jun 6, 2015 at 21:19

2 Answers 2


I would recommend a forward inclined centrifugal system, such as a forward curved fan.

The power input/output of any device, where fluid comes in/leaves with fluid rate $Q$ at $V$ and enters/exits at an angle $\theta$ at velocity $U$, would be:

$$\mathcal P = (V-U)(1-\cos(\theta))\rho QU$$.

If you have this device compress the gas, the power input runs in reverse. In both cases the angle helps. See the velocity triangle.

The real heart of this will comes down to putting some good valves on the openings. $U$ is a double edged sword - while it ups your power, if $V$ isn't very high compared to $U$ nothing is really happening. Don't forget $Q$ depends on $V$ or $U$, depending on how you look at it. The key to modifying this is to throttle your inlet opening down (whichever way you run) to a very small opening to have the highest $V$ possible, while keeping the outlet carefully controlled to not constrict $Q$ or $U$ beyond what is necessary to keep $V/U$ decent.

Perhaps using this as a first stage in a two stage rotary compressor could also help - the second stage is a true rotary compressor to really boost the pressure, but this assists the second stage to increase pressure beyond atmospheric.

Ultimately no device on the market will be built to this strange service - but by having a fairly symmetric rotary system with carefully controlled inputs should yield some decent results. I would definitely consult with a custom fan manufacturer.

  • $\begingroup$ Your thoughts seem to reinforce my assumption that these turbomachines have the necessary characteristics to be reversible. A forward inclined fan looks extremely similar in operation to a centrifugal compressor; is a centrifugal compressor also considered a forward inclined centrifugal system? The compressor, if the same as the fan, would appear to have a few advantages over the fan. And I see what you mean about carefully controlling the different variables in the system. That is definitely necessary for this application. The equation is especially helpful $\endgroup$
    – MikeJava
    Commented Jul 10, 2015 at 18:24
  • $\begingroup$ In theory that equation goes forwards or backwards. However, when using the same device, since $Q$ is based upon the cross sectional area, V and U are optimized for one way in the design of the inlet and outlet. Your best bet is to have it optimized for generation, since you've only got a limited container of pressurized air, but theoretically a limitless supply to compress. $\endgroup$
    – Mark
    Commented Jul 10, 2015 at 18:36
  • $\begingroup$ That's all very well when viewed in the context of the theory of a reversible centrifugal fan, but the very classification 'fan' implies that it is not suited for raising gas to a very high pressure at all. Can a centrifugal fan get air up to a respectable psi? $\endgroup$
    – MikeJava
    Commented Jul 15, 2015 at 2:46
  • $\begingroup$ I would say it wouldn't get to a high psi. That's why I'd recommend it as a first stage. I'm also using "stage" in a broad context - like the first few layers of the rotary compressor are shaped differently than the rest. So you would pump the compressed air back into the middle of the compressor (i.e. the last of those different shaped layers) not the very last layer of the device. $\endgroup$
    – Mark
    Commented Jul 16, 2015 at 1:06

I'm not an expert on air compressors or motors, but from my limited knowledge I think as you say that the centrifugal compressor would be the best for compression and a Tesla Turbine would be ideal as the motor. I think it should be possible to mount them on the same shaft but in separate airtight chambers, with some valves such that when the turbine is in operation the air is pumped out of the compressor chamber, and vice versa, so as not to cause undue resistance from the other impeller. Alternatively a clutch/grip mechanism that selects which one turns with the shaft.

Such a device could be considered a both-ways compressor/motor. To try to do it both ways with an impeller optimised for one of those scenarios seems like you will always be inefficient in the other.

  • $\begingroup$ I certainly see how a good (although not completely elegant) solution can be reached by putting two different machines on the same drive shaft and using some fancy connecting pieces. That seems like the real-world-practical way that this could go. A design embracing this principle but using a vaned compressor and motor could also be aesthetically and functionally pleasing. It's interesting, though, that you bring up the tesla turbine. When I look at that, I see not only a turbine but also the basic design of a centrifugal compressor. Both of these things without blade angle to worry about. $\endgroup$
    – MikeJava
    Commented Jul 25, 2015 at 1:09
  • $\begingroup$ I did think the same thing about the Tesla turbine, but I couldn't find much information about how good it is as a compressor. I would be interested to see if anyone has tried to measure how good one is in both directions. $\endgroup$
    – jhabbott
    Commented Jul 25, 2015 at 17:30
  • $\begingroup$ The tesla turbine can indeed be used in The opposite direction. In this case it seems that it is referred to as a tesla pump. The only difference between the turbine and the pump appears to be that in one case, air is causing the rotors to accelerate, and in the other, he rotors are causing the air to accelerate. When the air acts on the rotors, it spirals inward. When the rotors act on the air, it spirals outward, increasing in velocity, exactly like a centrifugal compressor. The question then is, 'would a blade-less, centrifugal compressor work?' If the answer is yes, we may have our answer. $\endgroup$
    – MikeJava
    Commented Jul 25, 2015 at 21:00
  • $\begingroup$ Also I assume the centrifugal/centripetal trade off would apply to a vaned centrifugal compressor as well, not just the tesla turbine. $\endgroup$
    – MikeJava
    Commented Jul 28, 2015 at 20:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.