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/)
-   -   Epsilon and mixing length value for airfoil simulation (http://www.cfd-online.com/Forums/openfoam-solving/58019-epsilon-mixing-length-value-airfoil-simulation.html)

gian_ing January 29, 2009 07:50

Hi all, I'm trying to simulat
 
Hi all,
I'm trying to simulate a turbulent flow around an airfoil. I have some difficulties in evaluating the correct value for epsilon.
In particular I don't know which value of the mixing length I should use.
The domain is 3 meters long and 2 high.
The airfoil is located at 1 meter from the inlet and in the middle of the height, the chord length is 0.1 m, the speed is 100 m/s, Re=1e6.
Thank you in advance for your suggestions.

Bye
Gianluca

gian_ing January 29, 2009 09:33

Hi, this is what I obtained so
 
Hi, this is what I obtained so far.
I attach the directory of the case and a screenshot.
Thanks again.

http://www.cfd-online.com/OpenFOAM_D...es/1/10872.jpg

http://www.cfd-online.com/OpenFOAM_D...hment_icon.gif airfoil_simulation.zip

Gianluca

wolle1982 January 29, 2009 12:32

hi, your U value seams to b
 
hi,

your U value seams to be 10m/s instead of 100. you also will have to raise the size of your mesh to 40-100 times the chordlength.

antonio_ing January 29, 2009 15:17

Actually, Gianluca and I tried
 
Actually, Gianluca and I tried to do a different simulation with the k and epsilon of a simulation of prof.Jasak on a naca 0012. with the same chord length and the same distance. but you see the result (it is embarassing :-) )

maruthamuthu_venkatraman January 30, 2009 17:32

Hello Antonio and Gianluca,
 
Hello Antonio and Gianluca,
Try to keep y+ approx 1 to 5 and use sst k-w model. You shall check the best practise guidelines in ERCOFTAC for such flow.

Naca profile example is there in it... May be it helps you.

good luck

ivan_cozza January 31, 2009 14:15

Mmmm, the k-Omega SST coded
 
Mmmm,
the k-Omega SST coded in OF is based on a wall functions approach, so it should be necessary to use an y+ of about 30...
egards, Ivan

antonio_ing February 1, 2009 04:55

thanks Ivan and Maruthamuthu.
 
thanks Ivan and Maruthamuthu. We find out our bug :-). In the boundary file we did not specify that the profile surface was of wall type. There are other types in OpenFOAM (inlet, outlet, etc...)

Is it so important to specify the other types (apart wall of course)?


All times are GMT -4. The time now is 23:04.