# Number of nodes for LES And DES

 Register Blogs Members List Search Today's Posts Mark Forums Read

 May 14, 2011, 02:41 Number of nodes for LES And DES #1 Senior Member   mohammad Join Date: Dec 2010 Location: UK Posts: 245 Rep Power: 16 Hey guys, I have heard that there are some quantitative equation( as rule of thumb) between number of nodes and "Re" number for DES and LES methods. Do you know anything about this matter?

May 14, 2011, 23:33
#2
Senior Member

Vieri Abolaffio
Join Date: Jul 2010
Location: Always on the move.
Posts: 308
Rep Power: 17
Quote:
 Originally Posted by mohammad Hey guys, I have heard that there are some quantitative equation( as rule of thumb) between number of nodes and "Re" number for DES and LES methods. Do you know anything about this matter?
it vary depending on the type of flow that you are investigating, for example LES of free shear flow, upon a certain limit does not vary significantly (re^0.8), while for wall bounded flows, with fully resolved viscous sublayer (no wall function), it varies with the Re^2.4

DNS for the same flows is around Re^3 and Re^3.6

you can obtain those figures using the definition of kolmogrov lenghtscale for the free shear and the friction velocity for the wall bounded flow.

May 14, 2011, 23:53
#3
Senior Member

Join Date: Dec 2010
Location: UK
Posts: 245
Rep Power: 16
Quote:
 Originally Posted by sail it vary depending on the type of flow that you are investigating, for example LES of free shear flow, upon a certain limit does not vary significantly (re^0.8),
Dear sail,
My case is a NREL wind turbine and the average Re in my case is between 500,000 to 900,000( the Reynolds is calculated locally according to the local air velocity and chord).
The average chord is 0.5 m and the blade span is 5m. I've put 53 nodes in chordwise and 83 in spanwise directions and based on CFX Post results my Y+<1. I paid attention to have finer mesh near the blade and further away mesh is coarse. the total number of nodes are about 5. million.

for the case above, do you have any suggestion about the number of nodes I need ?

Thanks a lot

May 15, 2011, 00:13
#4
Senior Member

Vieri Abolaffio
Join Date: Jul 2010
Location: Always on the move.
Posts: 308
Rep Power: 17
Quote:
 Originally Posted by mohammad Dear sail, My case is a NREL wind turbine and the average Re in my case is between 500,000 to 900,000( the Reynolds is calculated locally according to the local air velocity and chord). The average chord is 0.5 m and the blade span is 5m. I've put 53 nodes in chordwise and 83 in spanwise directions and based on CFX Post results my Y+<1. I paid attention to have finer mesh near the blade and further away mesh is coarse. the total number of nodes are about 5. million. for the case above, do you have any suggestion about the number of nodes I need ? Thanks a lot
oh, in those terms, you mean. no, can't give you an answer sorry.

but 50 pts chordwise and 83 spanwise looks a little bit coarse, have you done mesh indipencece studies yet?

just as a rule of thumb:
y+ < 1 ok
and also x+ < 100
and z+ < 20

May 15, 2011, 00:31
#5
Senior Member

Join Date: Dec 2010
Location: UK
Posts: 245
Rep Power: 16
Quote:
 Originally Posted by sail have you done mesh indipencece studies yet?
Actually I'm doing mesh sensitivity test now. I'm doing it for steady state case using SST model. At the same time i'm changing No. of nodes to make my model ready for DES or LES. What I have got till now is that DES is more efficient than LES for my case( regarding the time and machine power). but still i'm trying to understand the least adequate no. of nodes for my case for DES. If my model needs very large number then I should think about using a better PC.

Quote:
 Originally Posted by sail and also x+ < 100 and z+ < 20
which one is chordwise direction here? z or x? plus, how to calculate the values of x+ and z+ manually? Are they available in CFD result file?

 Tags des, les, number of nodes