result on airfoil wortmann fx63-137 simulation

jdfortune · January 29, 2011, 05:29

Hi,
I am doing a simulation on this airfoil to obtain the similar aerodynamic curve as from the literature
http://www.worldofkrauss.com/foils/243

the maximum Cl is obtained at correct angle which is at 11 degree. However, the value is slightly off. The value obtained from simulation is
Cl=1.7 while real Cl=2.013 from the website
Cd=0.18 which is much higher than real Cd=0.05

Im using spalart allmaras and Re=1 million
any advice is appreciated thanks

attached is the Y+ plot on airfoil wall.
And here is the url to download my mesh file
http://www.sendspace.com/file/tld05b

DoHander · January 29, 2011, 10:35

You won't be able to obtain the correct Cd with Spalart Allmaras, this is because the model, as implemented in Fluent, assumes the entire flow domain is turbulent which is not true.

For a short explanation check this article:

http://larcase.etsmtl.ca/PDF/Note-Paul-f.pdf

Do

DoHander · January 29, 2011, 10:36

In Fluent 12 and 13 you can use a transition model, this will give you a better Cd.

jdfortune · January 30, 2011, 11:47

thank you for the input DoHander, very much appreciated.
I've read the paper you gave and it seems that fully turbulent model on the entire domain is the reason for overprediction of Cd. So the overprediction is due to the skin friction alone or both pressure and friction drag?

I'm using Fluent 6.3, so does fluent 6.3 has transition model?
Or do i have to shift to version 12/13?

How about the error in Cl (around 15%)? is it due to the meshing not being fine enough?

thanks.

swiftaircraft · January 31, 2011, 17:17

Also your y+ values are very high. For accurate predictions you need to resolve the boundary layer. You will not be doing that with the mesh you currently have.

DoHander · January 31, 2011, 20:26

@jdfortune Fluent 6.3 doesn't have any transition model, you will need to switch to Fluent 12 or 13 if you want to use a transition model.

Also your mesh is bad, as swiftaircraft has noticed. If you want good results for Cd y+ must be less than 1 OR from 30 to 100. You can't use a mesh with y+ from 1 to 30 !!!

My advice is to try to keep your y+ from 30 to 100, this will allow you to use a reasonable mesh (from a memory usage perspective).

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
NO STAGNATION POINT FOR AIRFOIL SIMULATION	Rif	Main CFD Forum	6	February 4, 2008 08:33
Airfoil Simulation for Validation Purposes	Angela Bong	Main CFD Forum	7	September 13, 2006 14:04
how to make sure the simulation result is correct?	sham81	CFX	3	March 22, 2004 17:41
TASCflow simulation result problem?	Mason	CFX	0	February 22, 2004 08:54
Looking for :Multi-element airfoil simulation	Larry	Main CFD Forum	0	February 8, 2001 11:14

January 29, 2011, 10:35		#2
DoHander Senior Member Join Date: Nov 2009 Posts: 411 Rep Power: 20	You won't be able to obtain the correct Cd with Spalart Allmaras, this is because the model, as implemented in Fluent, assumes the entire flow domain is turbulent which is not true. For a short explanation check this article: http://larcase.etsmtl.ca/PDF/Note-Paul-f.pdf Do

January 29, 2011, 10:36		#3
DoHander Senior Member Join Date: Nov 2009 Posts: 411 Rep Power: 20	In Fluent 12 and 13 you can use a transition model, this will give you a better Cd.

January 30, 2011, 11:47		#4
jdfortune New Member jason fortune Join Date: Jan 2011 Posts: 4 Rep Power: 15	thank you for the input DoHander, very much appreciated. I've read the paper you gave and it seems that fully turbulent model on the entire domain is the reason for overprediction of Cd. So the overprediction is due to the skin friction alone or both pressure and friction drag? I'm using Fluent 6.3, so does fluent 6.3 has transition model? Or do i have to shift to version 12/13? How about the error in Cl (around 15%)? is it due to the meshing not being fine enough? thanks.

January 31, 2011, 17:17		#5
swiftaircraft Member David Stanbridge Join Date: Apr 2010 Location: Norwich, UK Posts: 59 Rep Power: 16	Also your y+ values are very high. For accurate predictions you need to resolve the boundary layer. You will not be doing that with the mesh you currently have.

January 31, 2011, 20:26		#6
DoHander Senior Member Join Date: Nov 2009 Posts: 411 Rep Power: 20	@jdfortune Fluent 6.3 doesn't have any transition model, you will need to switch to Fluent 12 or 13 if you want to use a transition model. Also your mesh is bad, as swiftaircraft has noticed. If you want good results for Cd y+ must be less than 1 OR from 30 to 100. You can't use a mesh with y+ from 1 to 30 !!! My advice is to try to keep your y+ from 30 to 100, this will allow you to use a reasonable mesh (from a memory usage perspective).