Fly into a Wind Tunnel with Autodesk Project Falcon

Sample wind tunnel test simulation in Autodesk Project Falcon showing air pressure around an airplane.

Running wind tunnel simulation in Autodesk Project Falcon to study wind velocity around an ariplane.

Project Falcon, Autodesk‘s latest technology preview at Autodesk Labs, is flying straight into the chambers where specialists conduct wind tunnel tests. The downloadable application functions as an independent software program, but it’s designed to complement surfacing programs like Autodesk Alias.

Once installed, Project Falcon allows you to import a design as an STL file, then simulate a wind tunnel test on it. The interface is fairly straightforward, with just enough controls to let you run basic tests. You set the wind velocity and direction, set the analysis plane’s orientation, specify the type of results you want to see (pressure, airflow, or displacement volume), then watch the screen come to live.

The results — wind flow patterns and pressure fields — are displayed in real time (or near real time speed, if your computer has slower processors). Specialists who wish to do more detailed FEA and CFD analyses may need more sophisticated packages, but for design engineers exploring different geometric shapes and curvatures, Project Falcon’s tool set is just the right amount.

If you wish to share your results with someone, or keep a record of it to defend your decisions, you may record the wind tunnel test results as AVI files and save them to your local drives. For a wind tunnel simulation program, the software is surprisingly lightweight. It runs on a typical workstation or PC (obviously, if you have multicore CPUs with larger memory footprint, you’ll get better performance).

In my view, digital wind tunnel simulation is not a replacement of physical wind tunnel test, but an easy-to-use software like Project Falcon can help steer you in the right direction in early conceptual phases. You can use it to avoid designs that have problematic geometry and curvatures so that when the physical prototype is ready to go through a real wind tunnel test, you can be confident it won’t produce catastrophic design flaws.

For more, watch the video demonstration below:

Update: I couldn’t resist the idea of running a wind-tunnel test on my own head, so I did it. Here’s the process:

  1. Take a series of 2D photos of my head from different angles.
  2. Upload them to Autodesk 123D Catch to produce a 3D scene.
  3. Save the head model as a mesh model in OBJ.
  4. Open the OBJ model in Autodesk 123D, save it again as STL.
  5. Run wind tunnel test on the STL model in Autodesk Project Falcon.

If you enjoyed this post, make sure you subscribe to my RSS feed!

26 Responses to Fly into a Wind Tunnel with Autodesk Project Falcon

  • Jim Watkins says:

    LOL Kenneth. Let’s see the wind tunnel test. Bald is beautiful in a wind tunnel. – Jim

  • Kenneth says:

    Jim, believe me, I’m very tempted to make it public. :-)

  • Thank you for your positive and thorough review of Project Falcon. Here is my head in a wind tunnel.

    http://labs.blogs.com/its_alive_in_the_lab/2011/08/project-falcon-for-alias-software-now-available.html

    Although I can’t change the shape of my head, I can cut my hair. This brings new meaning to the Crosby, Stills, Nash, and Young song” Almost Cut My Hair.”

  • Kenneth says:

    Aww, Scott! That looks great! I get much better insights into your head by observing the windflow patterns around it. :-) OK, you’ve inspired me. I think I will do a wind tunnel test on my shaved head and make it public.

  • José Solano says:

    Can i run this program whitout having Alias installed on my computer, Can I import a desing from Inventor to this program?

  • Kenneth says:

    Jose: Yes, you can run Project Falcon without having Autodesk Alias installed. From Autodesk Inventor, you can export your design as a STL file. This will allow you to run wind tunnel simulation on the design.

  • Ibrahim says:

    I’m trying to use this software to test different car designs for a school project, however I’m finding that the force of drag and drag coefficient both start at a relatively high value then decrease rapidly before leveling off (at about 10 seconds), is this normal?

  • Matt says:

    So is this a Navier-Stokes solver or a panel code? This description is really vague.

  • I had a good laugh but then again I am even balder my friend so I cannot laugh too hard. I wonder how good this program really is? In other words is is usefull enough to let others know about it? Good post Kenny. It is nice to get a laugh of two once in awhile. Thanks.

  • Matt says:

    I asked this once already, not more than 5 min ago, and the question was removed. So at the risk of sounding like a broken record, can someone please tell me if this is a Navier-Stokes solver or a panel code?

  • Matt says:

    Nevermind, it is ‘waiting moderation’… Sorry for the triple post.

  • Kenneth says:

    Ibrahim, Matt: Thanks for your comments! I’m not certain about the exact solver used in the application, and since I’m a CAD user and not an FEA expert, I’m probably not qualified to do a formal evaluation. I’d recommend posing your questions to Autodesk SIM Squad at http://usa.autodesk.com/adsk/servlet/pc/index?id=16661243&siteID=123112. I think they’ll be able to give you more information.

    Richard: I had a good laugh when I saw the results. I was already laughing when I saw my own head as a mesh model. Glad you liked the post!

  • Philip says:

    It would be interesting to build some models to test against known data! Maybe a Cessna for starters before moving into transonic territory?

  • Kenneth says:

    Philip, I agree. That’s one of the neglected areas of digital simulation. Many designers and engineers make critical decisions based on digital simulation results, but we seldom hear about the difference (if any) between what digital simulation predicts and how the product performs in real life under the same conditions.

  • Ibrahim says:

    I don’t think it will let you go at transonic speeds, the speed is capped at around 450 mph

  • Dennis Nagy says:

    Hi Kenneth,
    Very funny and very interesting! But, after the initial novelty/fun, the real key question emerges: is it “accurate enough” to guide engineers in the right direction? You said above “Specialists who wish to do more detailed FEA and CFD analyses may need more sophisticated packages, but for design engineers exploring different geometric shapes and curvatures, Project Falcon’s tool set is just the right amount.”
    Then later you said “…since I’m a CAD user and not an FEA expert, I’m probably not qualified to do a formal evaluation.” So how can you credibly claim that it is “just the right amount”? It would seem to me that a credible judgement of what is “just the right amount” would have to come from a fluid dynamics expert (or better yet, experts), not CFD or FEA experts (CFD and FEA are not mutually exclusive sub-segments of CAE/simulation).

  • Kenneth says:

    Dennis: That’s a fair point! By “just the right amount,” I meant to say the software gives a simple user interface with enough input fields to investigate and visualize how geometry and curvature changes in a design might affect the wind flow, drag, and lift of a vehicle or an aircraft. I can see that the approach taken by Autodesk Project Falcon would be less intimidating than some other specialized packages for people who are not simulation experts and specialists.

    Whether designers and engineers should be meddling with simulation and analysis software or leave them to experts and specialists is a good topic — one that I didn’t adequately address in the blog post above. I’d welcome more of your thoughts on it.

  • This is not a panel code. I don’t know enough about the precise physics to comment.

    By the way, you can get a wind tunnel for your iPhone: http://algorizk.com/windtunnel/

    I believe this is unrelated to Project Falcon.

    Keith

  • Disclaimer – I am a member of the Simulation team at Autodesk.

    Dennis’ question is valid. Regardless of the audience, the user expectation is the same. They expect the results to be fast and accurate. The idea behind Falcon is to be a part of a new generation of simulation tools. These tools are geared toward solving a specific problem at a specific point in the product development process. It uses many of the same numerics and algorithms that other commercial CFD codes do, but it is presented to the user in a different way. We tried to present the absolute minimal controls to give users a fast and accurate solution to a very specific problem. I think that is what Ken was referring to as “the right amount”. Internally, we have done a variety of validations on classical fluid mechanics problems, these are presented on the labs site mentioned above. In addition, we are doing more extensive testing to compare physical wind tunnel tests. At the end of the day, I agree with Dennis, accuracy is important. We are constantly collecting feedback and interested in what others have to say..

  • Kenneth says:

    Derrek: Thanks for the input!

  • Sermed Al-Wasiti says:

    Thank you for this article. I am a FEA / CFD user, though CFD is a small portion of my work, which is mostly in the vibration / stress analysis field.

    Is Project Falcon a “stand-alone” application, in the sense that it is similar to a concept car at an automobile show, developed to showcase some characteristics then not produced for the public ? Or is it the first step towards developing a CFD app like ANSYS’ CFX or Fluent ?

    It would be useful if Autodesk could develop its own CFD solution, to complement Inventor or some other MCAD app in the roster…

    On a lighter note, Kenneth, your head looks somewhat like a Klingon’s to me :D

  • Kenneth says:

    Sermed, thanks for writing! Project Falcon is “stand-alone” in the sense that it can be installed and used by itself. It’s not a plug-in. You might think of Project Falcon as an introductory app for wind tunnel simulation for FEA novices. Autodesk has its own CFD package, based on the technology it acquired from Blue Ridge Numerics. (http://www.deskeng.com/virtual_desktop/?p=4117)

    I may look like a Klingon, but I’m a Vulcan at heart. ;-)

  • Pingback: Finland Air Force - Page 4

  • Tabio says:

    Dear,

    I have tested airmodel in the Falcon, but I believe the drag results is a little higher than I expexted if compared if theorical equations.

    Is it possible I sent the stl file to you and we compare your analys with mine?

    Thank you

    Best Regards,

  • Marijan Pollak says:

    Hi, I tried to use FALCON for analyse of new kind of wind turbines and since it is BETA, and I am proffesional programmer and inventor, I would have lot of suggestions to make, but cannot contact Mr. Scott since published email is wrong.
    Anyhow, I hope Kenneth can read and help.
    Of late I have FALCON hanging and getting occupied by some internal work, so I wondered if there is newer version to test.
    My turbine is different from others because it use reactional force to produce rotation, and it does this by expelling fastened wind flow on turbine side at 90 degree angle to the original wind vector. Since air flow is at straight angle to wind flow, program see it as air standing still, regardless if it travel even 32 times faster than original wind flow (results from another CFD package). Therefore it would be usefull to see arrows of different sizes where size would visualize the volume of air in movement and head would point in actual movement direction. At present, arrows are flat and wide but looked from another angle they appear to be thin. Also, very rarely could be seen dark blue arrows traveling at 90 degrees angle to original wond flow, but displaced far behind turbine. It would be usefull to keep the Slice centered on object center all the time, and it would be still nicer to see turbine actually turning. If such things are added
    as measurement of torque force, RPM and similar, and if one can put weight on turbine to see how would load influence the rotation speed, it would be allmost complete.
    Next what I noticed that when looking tubine head on, i could see tornado forming in fronth of its mouth and air flow arrows zeroing in on turbine entrance, but this cannot be seen either in top wiew nor side wiew. “Head up” controls are not working at all, and it would be usefull to be able to see which exactly STL file is being tested. Sometimes on slope surfaces (as in the funnel) appear to have big tooths which then
    make obstacles to air flow and spoil the results. Sometimes object contours could be seen but most of the time not. One has to be quite careful not to change things fast
    as clicking on different controls sometime hang that control frame/window. If I want to
    read in another STL file, all painstackingly set parameters get reset to 2D concept instead being the same 3D as set. Last but not least, it would be nice if there could be
    a way to save settings of wind tunnel. As I am proffesional programmer I know plenty of tricks how to get more work from computer (from times when computers can do few thousand calculations per minute only) and I can help You to make FALCON fast as lightning, if it is written in some fully implemented OO language. As I planned to write
    my own 3D simulation program with all kind of testings done in paralel, I guess I could
    be usefull member of development team. Just imagine FALCON as 100 or 1000 times faster and more accurate than it is.
    Regards from Croatia, the homeland of Engineer Nikola Tesla!

    Marijan Pollak, IT SA/SE 1st. Class, Instructor and Team Leader,
    Proffesional Problem Solver, Ecconomist and Inventor

  • Pingback: Project Falcon Graduates to Become Autodesk Flow Design | Virtual Desktop

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>