Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Aerospace Wind Tunnel Design Question Re: Test Section and Diffuser/Fan Size

  1. Jan 14, 2013 #1
    My name is Ryan and I'm a volunteer instructor with the local Royal Canadian Air Cadet squadron. I'm looking for ways to try and make the kids' aviation lessons more interactive and hands on, and so I've decided I'd like to build them a small wind tunnel, purely for flow visualization, so that they can see and play with some of the principals they're learning about.

    I'm in the research stage, and I've done a ton of reading, but I haven't been able to find the answer to this specific question, so hopefully someone can give me a hand.

    Essentially, I want the test section to be as large as possible to allow for a multitude of models to be placed in it, but I also need the entire unit to be relatively compact. I'd rather avoid using a larger fan, so I'm thinking of using a 6" diameter electronics cooling fan that puts out about 179 cfm.

    My question is, if I'm using a 6" diameter fan, does the test section need to be equal to or smaller than that for the whole thing to work properly? I've yet to come across a photo of a tunnel where the test section was larger than the end of the diffuser section.

    Could I get away with say an 7" x 7" test section, with say an 8" x 8" max. diffuser housing the 6" diam fan? Or is there some fundamental aspect of wind tunnel design that requires the test section to be significantly smaller than the diffuser/fan?

    Any advice/help would be greatly appreciated.

    Ryan Healy
  2. jcsd
  3. Jan 15, 2013 #2


    User Avatar
    Gold Member

    Afaik, the constriction of the test section is aimed at getting the most uniform flow possible. If the source is a 6" fan blowing into a 7" section, that will be difficult. An 8" diffuser section will not really do much to even out the flow unless it is pretty long, which conflicts with your wish for a compact setup.

    If the objective is simply to visualize flow, a water tunnel might be an alternative, because while only 2D, it can carry dye plumes that really show laminar and turbulent flows very clearly. Smoke plumes in a small wind tunnel are much more difficult to set up and to observe.
  4. Jan 15, 2013 #3
    It is important that the test section diameter be smaller than the fan diameter. From a practical standpoint, the velocity of the air directly behind the fan will not be very fast. In fact, there is a lower limit to wind tunnel freestream test speeds before your results will be useless. While I do not know the justification for this fact, I believe it has something to do with large scale turbulent effects. In addition, using a diffuser to bridge the test section may create flow disturbances at the corners where the diameter begins to expand. For example you may have some separation.

    I agree with etudiant that a water tunnel will be much easier for your project. You can use simple food coloring to show some pretty cool flow phenomena. You can buy pumps that will have the required mass flow very cheaply. Just remember to check your Reynolds Number!
  5. Jan 15, 2013 #4
    Thanks for the feedback! Luckily I was able to find a 9" diameter fan DC cooling fan that moves about 159 CFM, and so I'm planning to use that one. My plan is now to go with a 9" fan, mounted at the end of a diffuser section 10" square at its max width. The test section will be 8"x8"x14", and the contraction cone will go from 16" at the mouth to 8" at the test section. I will include a settling chamber (not included in the diagram). Also not in the diagram are the flange joints that will be used to couple each major section (as I want it to be modular so it can be easily disassembled and should the kids break part of it, only that section will need to be rebuilt.)

    Major component rough dimensions:

    This approximates the proportions of the NASA Wright Memorial Tunnel, but at a smaller scale which will fit on an average teacher's desk. The Nasa WMT can be seen here:

    Does this seem like in theory it would be ok? Most of the homemade small scale tunnels I've seen made for flow visualization seem to operate in and around these kinds of proportions. Obviously this is hardly the scientific way to size things up, but does anything here stand out as a glaring problem?
  6. Jan 15, 2013 #5


    User Avatar
    Gold Member

    I have no experience with small tunnels, but wonder whether this arrangement would provide enough air to the fan. I'm pretty sure the fan rating is reduced by needing to draw the air through a four foot duct. It may still be quite adequate of course, but check your numbers.
    A 159 ft**3/min fan would at best move about 2.5 ft**3/sec. That suggests the flow through your roughly 1/2 ft**3 test section is about 5 ft/sec, about 3.5 mph. Can you get useful results at that speed?
  7. Jan 15, 2013 #6
    If I recall correctly from my internship I was working on either a 8x8 or 12x12" test section. Our minimum allowable speed was Mach .1 before our instruments could not record reliable numbers. That being said, we had a very small tunnel with about a 4x4" section which could go much slower. I think ideal speeds were 20mph.
  8. Jan 15, 2013 #7
    Since you are only doing qualitative demonstrations for kids I am sure this is fine. Depending on the type of flow visualization thou want to do slow speeds are better. What are you planning on doing?

    You will probably need to place some sort of flow straightener upstream of the test section. You can make a honeycomb screen out of drinking straws for example.

    I agree that a wate tunnel would be better for flow visualization however it would be harder to transport and build.
  9. Jan 15, 2013 #8
    More excellent feedback! Thanks a ton guys. Keep it comin' if you've got it!

    I trust your calculations more than mine. Based on a set of calculations I found on another forum, my CPM to MPH calculation goes as follows:

    (CFM / Sq. Feet Of Test Area) * 60 = fph

    fph / 5280 = mph

    That would work out to (159 CFM / .0556 sq feet) * 60 = 171582.72 fph
    171582.72 / 5280 = 32.49 mph

    That also doesn't take into account the velocity increase that should happen as the air goes through the contraction cone.

    That math seemed to be the agreed upon CFM to mph conversion on the other forum, but again, no idea if its really correct. Even if it is, 32.49 mph is much too fast for the purposes I need, so if that were the case, I'd be using a pot or variac to slow things down anyways.

    Does that not add up though?


    The visualizations I'm planning this to be used for are simple ones like flow patterns over different types of airfoils, streamlines, visualization of a stall caused by separation at high alpha, etc. Then on top of that, just for the kids to be able to do fun things like put scale model fighter jets in and stuff like that for smoke or tuft visualizations (this test section is designed to just barely accommodate a 1:72 F/A-18A/B/C/D Hornet).

    I'll definitely be including a settling chamber for flow straightening. I may go the straw bundle route, but I'll probably try to find a good honeycomb mesh first.

    As far as a water tunnel goes, dye in water would definitely show better than dry ice or smoke pellet smoke through a small win tunnel, but one of the main goals of this project is for it to be easy to transport in sections and require minimal setup and tear down. A water tunnel is just a bit more of a pain in the butt. Also, the "cool factor" of having their own wind tunnel is definitely a big part of it, since in the end this is all about finding more engaging and interactive ways for the kids to learn about aviation and aeronautics.
    Last edited: Jan 15, 2013
  10. Jan 15, 2013 #9


    User Avatar
    Gold Member

    Please correct me, but the sketch shows the test area is 8"x8", or 4/9th of a square foot, about 0.44 ft**2. rather than the 0.0556ft**2 shown above.
    The speed will be around 3.5mph given that the test cross section is much larger than the 0.0556 ft**2 in your calculation.

    I'm just assuming the 160 cubic feet/min will sweep the roughly half cubic foot test cell 320 times/min, or about 5.4 times/sec, so the air speed is about 5 ft/sec.
  11. Jan 15, 2013 #10
    You only want slow speeds if you are going to use a water tunnel. As I said earlier, if you want to use air then the flow has to be moving fast enough such that you can ignore the large scale effects of the flow. As an example, watch dust in your room (if you have a particularly dusty room) or even in a bathroom. Turn the bathroom fan on, notice that nothing really happens to the flow pattern in the dust except really close to the fan. If the fan had a sufficiently high mass flow, then the dust in the bathroom would most likely move towards the fan uniformly. I'm not an expert on wind tunnels, I am basing my explanation experience running a few different tunnels.

    Is this really so? I think this tunnel could be easily built out of PVC, especially if Ryan decides to go with a small design. Clear PVC should be available for the test section, diffusers/nozzles are readily available in any hardware store along with sealant is readily available. The only difficulty will be in implementing the motor. You could use a centrifugal fan, I know some small water pumps are sold as drill attachments. Just an idea.
  12. Jan 15, 2013 #11

    Sure enough! When I was using the sq. inches to sq. feet converter on Google (because dividing by 144 is extremely hard haha!) I was putting in 8 instead of 64!!!!:redface:

    So based on the CORRECT number (0.444 like you say), I get:

    (159 / 0.444) x 60 = 21486.49 fph

    21486.49 / 5280 = 4.07 mph.

    In theory, if my understanding is correct, this would be the speed in a straight tunnel. In theory, again, if my understanding is correct, the contraction cone should boost this speed somewhat. I think this might be ok for my purposes.

    One of the good things about building it in modular sections is that if this turns out to be too weak, I can just rebuild the diffuser/power section with a larger, more powerful fan.
  13. Jan 15, 2013 #12
    Another question I'm having trouble with:
    Is symmetry important in the tunnel design? Or is it less of a factor since the flow is straightened anyways?

    For instance, does the symmetrical design on the right have some sort of functional benefit over the asymmetrical design on the left? (Assuming all else equal including volume of the contraction and diffuser sections)


    The flat bottom would make construction easier and aid with stability and stress, which is important since this is meant to be used by kids (and potentially built by kids as a future project).

    But will the shape have any adverse effect on the flow compared to the symmetrical one?
  14. Jan 15, 2013 #13


    User Avatar
    Gold Member

    A speed of 10-20mph should be about right, much more than that and there would be no time to see the vortices really develop before they get flushed.
    A 500 cubic foot/min duct fan booster costs about $50 and should work. Not sure they can be speed controlled though, which may be a killer.
    It is not clear to me how the flows will be made visible. Probes to inject smoke, maybe from liquid nitrogen, need to be fine to minimize the flow disruption, but also need to be easily repositioned, so there is a need for access ports.
    It is not clear that this tunnel will be sensitive enough to allow the kids to see how differing models perform, but it should work for the basics. It would be a lot better than what the Wright brothers had when they worked the problem.
  15. Jan 15, 2013 #14
    The only reason why symmetric tunnel designs are used is because it is easier to obtain an uniform velocity profile. In your case you don't need to be concerned with this issue. I still think you should consider using PVC for your construction - it will save you time and money.
  16. Jan 15, 2013 #15
    The duct fan booster is a great idea! Saw a few 10" at 650cfm for around $60. As far as speed, I'm thinking if I hoop it up to a variable AC adapter that should do the trick for speed control.
  17. Jan 15, 2013 #16
    Good to know on the symmetry. PVC is an interesting idea. I've noticed that locally PVC sheets are a bit harder to comeby, and at larger thicknesses, are more expensive than sanded cabinet grade plywood. It'd certainly be a nice weight reduction though.

    My plan right now is to make the test section out of 0.220" Lexan polycarbonate, and to make the diffuser and contraction sections out of 1/2" birch plywood. The different sections will all be bolted together using flanges made of 1/4" plywood with gaskets fitted between them to stop air leakage.

    I'd love to do it with sheet metal, but I have no experience working it so I'll stick to materials I'm more familiar with. If I could find PVC sheets of a decent thickness at a decent price, I'd definitely be interested in them.
  18. Jan 15, 2013 #17


    User Avatar
    Gold Member

    Entirely agree.
    The whole paraphanalia of diffuser etc are I think luxury appendages that don't really add value.
    A simple tunnel of PVC tubing, maybe with a flow straightener section stuffed with straws at the beginning, a clear test section that allows the test article to be well observed and a duct fan to power the whole thing should work adequately. The square section arrangement is probably a bit better for taking pictures and even air flow, but requires more assembly work and will be heavier.
    Your choice.
  19. Jan 15, 2013 #18
    My original choice was actually to use PVC pipe with a couple of windows cut out and some lexan over top, but the largest diameter I could find readily available was 4", which is smaller than I'm willing to use because in a classroom full of kids, it just wouldn't provide enough visibility.

    I'll talk to some plumbing supply stores and see if I can get my hands on some 8" or 10" PVC pipe. Shoving a duct fan booster in one end and a flow straightener in the other would sure be a hell of a lot easier than building out what I'm planning now.
  20. Jan 15, 2013 #19


    User Avatar
    Gold Member

    If PVC pipe is too hard to get in the right diameter, maybe just use ducts.
    I don't think that the flow would be disrupted if you made some large cutouts in it and bent thin plastic sheet to cover it. Even the plastic from loose leaf separators would give an 8x11 inch window. You could then just turn the model 90 degrees to show the top of the flow as well as the side view.
    Please keep us posted.
  21. Jan 15, 2013 #20
    Speaking of ducts..why not air conditioning ducts! They come in all shapes and sizes! Also, if you really wanted to you could heat the PVC then form it.
  22. Jan 15, 2013 #21


    User Avatar
    Science Advisor
    Homework Helper

    I would be careful about taking too many "short cuts" because "this is only demos for kids". First, because demos that don't work properly can be worse than useless, and second because (as the OP obviously already knows) if the Canadian air cadet organization is anything like the UK one, some of those "kids" will probably end up on the fast track to becoming air force pilots.

    The NASA site you linked to has detailed plans for a few different designs. They maiy look more complcated that you hoped for, but all that stuff is there for a reason. And after a while you will probably want to progress from qualitative demos to actually measuring something...

    There are good reasons for wanting to make something small and portable, but (within the limits of common sense) bigger, and higher air speeds are easier to use and more capable. If nothing else, if you can't use hand tools with both hands insde the working section of the tunnel, you are making it hard for yourself setting up experiments and demos!
  23. Jan 15, 2013 #22
    I'm looking into some large diameter PVC pipe. I'll make some calls tomorrow. If I can get 10" or 12" diameter pipe, then the booster fans make a very nice fit and a very good CFM compared to the fan system I was planning to use. Unfortunately, the downside is that large diameter pipe only seems to be sold in 10' lengths. I don't need nearly that much and the expense is significant.
  24. Jan 15, 2013 #23
    Certainly a good point. It's a bit of a catch 22 though because the Cadets meet weekly at a school, not at their own permanent location, meaning the tunnel will need to be brought to cadets and assembled when needed, thus it absolutely has to be portable and modular. Otherwise, I can build it as big as I want, but the cadets won't be able to use it.

    That being said, the design I've done the most planning around is essentially the Baals wind tunnel as demo'd on that NASA site. The only difference being that mine will be smaller, and the fan system will be smaller. The modular design I'm going for allows me to test that, and then if the larger fan is absolutely necessary, I can remove the old diffuser/power module and build in a new one with a larger, more powerful fan.

    As it stands though, for pure flow rate, the large diameter pvc pipe driven by the duct booster fan seems to be by far the best option. I'll have to continue my research and decide on one method or the other once I've got some more complete info on sourcing the large diameter pipe.

    Thanks again to everyone for all the great help this far! I really appreciate what's been offered up, and please, post any other thoughts that come to mind!
  25. Jan 16, 2013 #24


    User Avatar
    Gold Member

    The construction and utility companies often have that size pipe on their jobs and after cutting to fit the unused ends are discarded. That might solve both the length and the cost problem.
    Air conditioning ducts come in both rectangular as well as round cross section. They are cheap, light and easy to work with.
    AlephZero is entirely right that you need good access to the test section, else the mounting of the models is a chore. Maybe just cut the tunnel in half at that point with a hinge, so it opens like a clamshell.
    Presumably the NASA documents give some guidance on what instrumentation to consider and how to accomodate them, measurement is the soul of engineering and the kids should get that experience early.
  26. Jan 16, 2013 #25
    Will you be trying some sort of surface flow visualization such as fluorescent oil flow? If so, these techniques may not work well at very low speeds because the shear stress is not large enough.

    Keep in mind that if your model is too big relative to your test section then you can have some pretty serious wall effects that will alter behavior of your model. These effects can be severe enough to qualitatively change the flow so it is no longer representative of what the flow should be.

    Along the lines of what Alephzero said about taking too many short cuts, you should use the symmetrical inlet to avoid significantly skewing your flow in the est section.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook