CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM Running, Solving & CFD (http://www.cfd-online.com/Forums/openfoam-solving/)
-   -   naca 0012 validation case: wrong Cd (http://www.cfd-online.com/Forums/openfoam-solving/75019-naca-0012-validation-case-wrong-cd.html)

vaina74 April 14, 2010 13:04

naca 0012 validation case: wrong Cd
 
2 Attachment(s)
Since I was simulating an external flow on a blade section of a marine propeller and lift and drag coefficients seem to be good (I checked them with Xfoil), I thought to test a NACA 0012 airfoil. Setting are:

Re = 2E+6
\alpha = 1

According to experimental results, I obtain a good C_L=0.11 and a bad C_D=0.011 (vs 0.0064). I guess it depends on my fully turbulent models (\kappa-\omega and Spalart-Allmaras), because my coefficients are similar to Xfoil ones (if I force transition at x/c=0.1 on both surfaces). But I guess if I can improve my settings (BC, fvSchemes, fvSolution): any suggestions?

Meanwhile I'll go on with different angles of attack.

Marta September 28, 2010 09:45

Dear Vaina, have you found out something concerning the comparison between the Cl and Cd calculated by OFoam and those coming from experiments?
I am simulating the flow around a Naca airfoil too, but the final results are not as good as those of other CFD codes, even if the simulation has converged (residuals lower than 1e-6).

Moreover, i had to choose the Spalart-Allmaras equation for turbulence modelling, because with k-epsilon model i didn't succeed in completing the run.

Have you obtained something interesting or useful out of your trials?

Thank you very much,

Marta

chriss September 28, 2010 11:36

Cd
 
Quote:

Originally Posted by vaina74 (Post 254614)
Since I was simulating an external flow on a blade section of a marine propeller and lift and drag coefficients seem to be good (I checked them with Xfoil), I thought to test a NACA 0012 airfoil. Setting are:

Re = 2E+6
\alpha = 1

According to experimental results, I obtain a good C_L=0.11 and a bad C_D=0.011 (vs 0.0064). I guess it depends on my fully turbulent models (\kappa-\omega and Spalart-Allmaras), because my coefficients are similar to Xfoil ones (if I force transition at x/c=0.1 on both surfaces). But I guess if I can improve my settings (BC, fvSchemes, fvSolution): any suggestions?

Meanwhile I'll go on with different angles of attack.


Attention to its trailing edge. You need to trick the fluent little refining this site. Using k-omega sst results are good. If u want, i have some results.

Simon Lapointe September 28, 2010 11:37

It is normal that the drag coefficient computed with turbulent RANS models (as Spalart-Allmaras) be higher than the experimental or XFOIL values since the models predict a fully turbulent boundary layer but in reality there is an important region of laminar boundary layer. If you want a better prediction of Cd in general CFD codes, you'll have to use transitional RANS models.

fung_wern October 22, 2010 03:40

Hi can I know what source of validations you all use? Can you send the database to me please?
fung_wern@hotmail.com
Thank you!

aerothermal July 15, 2011 11:50

Hello Vaina,


DO you have your case available for download? I am interested in both system and constant folders, including mesh.

You mentioned:
Quote:

if I force transition at on both surfaces
What transition model are you using?

Regards,

Guilherme


All times are GMT -4. The time now is 18:57.