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/)
-   -   Trouble of LES case (http://www.cfd-online.com/Forums/openfoam-solving/60272-trouble-les-case.html)

tadashi March 13, 2006 16:11

Dear all In my case of chan
 
Dear all

In my case of channelOodles, the result of velocity seems laminar profile(parabolic profile) even at turbulent condition.
I think something my setup is wrong, but I can't find it.
Please help anybody!

Values of my case model are below.

Re number=30000.

turbulenceProperties are
LESmodel: Smagorinsky
delta : vanDriest

Values of Ubar and nu are written on transportProperties file.

Y+ values are below 10.

initial value of k is zero.
inlet outlet conditions are cyclic.
Other conditions are refered from tutorial sample.

Best regards

Tadashi

hjasak March 13, 2006 17:42

You should not start with init
 
You should not start with initial k = 0. Estimate uPrime to several percent of uBar and do

k = 3/2 uPrime^2

as the initial guess of k.

Hrv

tadashi March 14, 2006 08:52

Thank you very much Hrv. I wi
 
Thank you very much Hrv.
I will try again with your advice.

Tadashi

eugene March 14, 2006 09:32

What is your initial value of
 
What is your initial value of U?

The case will take ages to go turbulent from a uniform velocity unless you give it a push.

tadashi March 14, 2006 10:12

Hi Eugene Initial value of
 
Hi Eugene

Initial value of U is as same as value of Ubar for uniform, and I could get turbulent condition with my case. But the result of mean velocity profile is quite different from that of RANS result.

Is there any other value that I must check?

Tadashi

eugene March 14, 2006 11:57

Hmm, above you said your U pro
 
Hmm, above you said your U profile was parabolic. The only way for that to happen is for the resolved flow to be laminar (at least in my experience).

Could you post a picture of the instantaneous U and k in the channel?

tadashi March 14, 2006 12:52

Please refer as below. U \
 
Please refer as below.

U
\image(/home/tadashi/temp/LES_U.jpg)
k
\image(/home/tadashi/temp/LES_k.jpg)

tadashi March 14, 2006 13:48

Sorry, I am not good at these
 
Sorry, I am not good at these kinds.

U
http://www.cfd-online.com/OpenFOAM_D...your_image.gif
k
http://www.cfd-online.com/OpenFOAM_D...your_image.gif

tadashi March 24, 2006 14:12

Dear Eugene Sorry, I couldn
 
Dear Eugene

Sorry, I couldn't upload pics.
But I thought that the problem is cyclic boundary condition will just loose kinetic energy every iterations.
Finaly, flow will become laminar.
How to give kinetic energy to flow every iterations?
Or is my understanding wrong?
Thanks.

eugene March 25, 2006 10:00

channelOodles already takes ca
 
channelOodles already takes care of this via a mean pressure gradient that is adjusted to maintain the target mean velocity, Ubar, that is set in the transportProperties dictionary. Of course running normal oodels on a periodic flow will cause it to wind down.

Check the formatting section of the forum on how to upload pictures.


All times are GMT -4. The time now is 02:23.