I am a high school student learning to analyze aircraft models in Autodesk CFD for the first time. My current project is analyzing a supercritical airfoil on Autodesk CFD (that was designed in Autodesk Fusion360).

Can I get help with the following : I researched several turbulence models(https://knowledge.autodesk.com/support/cfd/learn-explore/caas/CloudHelp/cloudhelp/2019/ENU/SimCFD-UsersGuide/files/GUID-E9E8ACA1-8D49-4A49-8A35-52DB1A2C3E5F-htm.html#:~:text=Turb.%20model%20%20%20Turbulence%20Model%20%20,model%20wi%20...%20%206%20more%20rows) and I chose the SST k omega turbulence model, with a 250 m/s inlet velocity, 0 gauge pressure , unknown outlet and a compressible flow , I am getting a lift to drag ratio of approx. 4.3.I think the lift to drag ratio is too low. I am suspecting the analysis conditions are to blame. Are these the correct input parameters? Am I missing anything else? Thank you. Here is a screenshot of the analysis:I ran the airfoil analysis at angle of attack=5degrees
The following image is my mesh along with the convergence plot after working with the suggestions. This is my mesh along with the convergence plot after working with the suggestions.

  • $\begingroup$ You would probably do better to post this question on the Aviation SE site. $\endgroup$
    – Eric S
    Mar 12 at 14:52

Welcome to Engineering.SE

Regarding your question, in CFD analysis so many things can go wrong. You have a good start with some right inlet and initial conditions, but the problem you're simulating isn't an easy one, and requires taking care of so many details.

1 - Let's start with the boundary conditions enter image description here

If you're simulating a wind-tunnel-like situation or a free stream over a body, you should not keep the non-slip conditions for the walls (except the airfoil of course), this will affect your results greatly and it's adding the complexity of non-developed internal flow, so we need to avoid that. I am not expert in AutoCAD CFD, but you need to make upper and lower walls as "slip" walls.

2 - Convergence enter image description here Judging from the curves of residuals, I don't think you've achieved numerical convergence, so I wouldn't count on any of the results you obtained. A typical converged solution curves would look similarly like this: enter image description here

3 - Geometry of the problem is too small

Box-wind-tunnel simulations can give quite good results, but the typical 2D airfoil set up would look like the following picture, with a freestream $U_{\infty}$ boundary condition or simple inlet-outlet setup. Also there are the mesh refinements around the airfoil (all of the simulation "drama" happens right there!).

And, you did not show us your final mesh, so that can also be one of your simulation problems.

enter image description here

4 - Turbulence

The choice of $\text{SST} \ k-\omega$ is good, but what about the wall treatment? what is the max $y^+$ around the airfoil? Flow around airfoil involves adverse pressure gradients, so I would make my $y^+$ as low as it takes to capture the viscous layer.

You also should try different turbulence models, and in your case I would give Spalart–Allmaras model a try.

I hope I haven't scared you off, please investigate the above points we discussed, and let me know if anything isn't clear to you.

  • $\begingroup$ Thank you for all the detail. I genuinely appreciate it . (1)Boundary conditions: I edited the upper and lower walls to "slip walls" as you suggested. (2) Convergence: I am tempted to stop the analysis when the force readings don't change much from iteration to iteration( typically I run 500 iterations). I will post the new screenshot of my analysis as an edit to my original question. Do I need to run further iterations? Am I missing anything? (3) Mesh: The new screenshot shows the mesh I am currently using. Please comment. $\endgroup$
    – RoseDuda
    Mar 20 at 10:07
  • $\begingroup$ (4)Turbulence: I don't have Spalart-Allmaras model in Autodesk . So I tried the SST k omega DES(Detached Eddy Simulation) . I set my max y+ value to 20. How do I know if this number was sufficient to capture the viscous layer? $\endgroup$
    – RoseDuda
    Mar 20 at 10:14
  • $\begingroup$ @RoseDuda If you found the answer helpful, give it a vote or accept it so the system knows the question has been answered. $\endgroup$
    – Solar Mike
    Mar 20 at 12:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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