CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   SU2 (http://www.cfd-online.com/Forums/su2/)
-   -   Incompressible viscous flow around NACA0012 (http://www.cfd-online.com/Forums/su2/111610-incompressible-viscous-flow-around-naca0012.html)

robyTKD January 10, 2013 19:21

Incompressible viscous flow around NACA0012
 
Hello to everyone,

I am a Politecnico di Milano student and I am using SU2 in my Master thesis involving racecar aerodynamics.
My first goal is to analyze the viscous flow around NACA0012 profile in incompressible conditions; in particular I'm trying to reproduce these results: http://turbmodels.larc.nasa.gov/naca0012_val_sa.html.

The mesh used in my simulations is not the same as the one used in the report. It's a C-mesh of quadrangle elements around a NACA0012 of unitary chord length; boundaries of the domain are 30 chord unities far from the leading and trailing edge.

Unfortunately I am not able to make the solution converge; the only settings that gave me results are reported in the configuration file attached below, but they don't match the reference.
I've not been able to attach the mesh file because it exceeds maximum dimensions.

Could you please provide me help and feedback about my situation?

Thank you very much in advance.

Roberto

robyTKD January 10, 2013 19:31

1 Attachment(s)
And here there is the configuration file.

fpalacios January 18, 2013 00:46

1 Attachment(s)
Quote:

Originally Posted by robyTKD (Post 401290)
Hello to everyone,

I am a Politecnico di Milano student and I am using SU2 in my Master thesis involving racecar aerodynamics.
My first goal is to analyze the viscous flow around NACA0012 profile in incompressible conditions; in particular I'm trying to reproduce these results: http://turbmodels.larc.nasa.gov/naca0012_val_sa.html.

The mesh used in my simulations is not the same as the one used in the report. It's a C-mesh of quadrangle elements around a NACA0012 of unitary chord length; boundaries of the domain are 30 chord unities far from the leading and trailing edge.

Unfortunately I am not able to make the solution converge; the only settings that gave me results are reported in the configuration file attached below, but they don't match the reference.
I've not been able to attach the mesh file because it exceeds maximum dimensions.

Could you please provide me help and feedback about my situation?

Thank you very much in advance.

Roberto

Hi Roberto,

Thanks for your interest in SU2. I think that this simulation should work with the appropriate set of parameters. Please find attached a suggested config file, unfortunately I don't have the right grid for Reynolds 6M. Could you please post the grid that you are using?

Best,
Francisco

Attachment 18267

robyTKD January 18, 2013 11:23

3 Attachment(s)
Hi Francisco,

thank you for the reply and for the configuration file.
Unfortunately it didn't make the case converge, so I changed some settings and now it seems to converge(modified config file is attached below).
At 0 angle of attack the solution is right.

Moving to my case (10 degrees angle of attack), results of the simulation (Cl=0.99, Cd=0.022) differ from the reference (Cl=1.09, Cd=0.012); this is strange because the solution seems smooth and also the Cp is regular (see figures attached below).
How can I solve this problem? Is there an error in the code or my settings are wrong?

I have a further question: is it possible to run the ADJOINT with the incompressible formulation?

This is my mesh: http://www.4shared.com/zip/8Zgpycw1/...12_40ksu2.html.

Thank you very much,
Roberto

shirazbj February 6, 2013 03:03

2 Attachment(s)
I am interested in imcompressable case too.

I tried you cfg and mesh.

Here is a result after 2000 steps. Your grid may not suit.

dtucker April 21, 2013 13:10

Incompressible AoA??
 
Is there a way to specify AoA with an incompressible case without having to do the trig??

...call me lazy, but I want to do a large range of angles and it would be much faster if I could just specify a velocity and angle.

Is there a way to do this, or is there some obstacle I'm not considering?

Thanks,
Dave

taxalian April 21, 2013 13:38

Hi Roberto,

To better predict the cl values vs Experiment at higher angles of attack. Well i think, first of all you should make sure that your yplus value should be less than one.

Regards.


All times are GMT -4. The time now is 07:16.