|
[Sponsors] |
December 29, 2003, 03:20 |
LES
|
#1 |
Guest
Posts: n/a
|
HI All, I just want to have some guidance from experts here regarding LES in star CD. What are the criteria required for this model, like y+ or any other? Can I just switch on LES without changing the mesh in my earlier model where I used High Re K-E model? Thanks in advance....
|
|
December 31, 2003, 06:27 |
Re: LES
|
#2 |
Guest
Posts: n/a
|
LES does not use wall functions since it is not meant to. Thus, if you need to run a LES simulation, you should use a Low-Re mesh, i.e. with Y+ values below 1. Usually, you will require a much finer mesh than for RANSE simulation. Typical mesh sizes are in the order of several millions control volumes. Alternatively, you can use DES which uses a RANSE solution in wall vicinity and LES in the far-field. The only problem is that, when using RANSE, you assume a turbulent flow everywhere, and therefore, if you want to capture a laminar/turbulent transition, what LES will capture if the mesh is fine enough, you will never be able to get it
|
|
January 8, 2004, 13:36 |
Re: LES
|
#3 |
Guest
Posts: n/a
|
hi; The way LES is implimented in STAR-CD it is possible to use the high Reynolds k-e mesh (3D of course). Which means you are using a log-law wall function. When the grid is refined enough Y+=1, then the wall function is switched off. Remember to turn on the Vandariest damping function on (by setting Switch20=1). It is a bit tricky the way you need to save the geom and prob files (read the realise notes for information on this). For information, even a full LES will have problems in the transition. So let hope u case is not in this range of flows. good luck;
|
|
February 5, 2004, 13:19 |
Re: LES
|
#4 |
Guest
Posts: n/a
|
Hi Yacine, I have a question for you: where I find the realise notes? Thank you in advance.
|
|
Thread Tools | Search this Thread |
Display Modes | |
|
|