CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   HELP Please! (http://www.cfd-online.com/Forums/fluent/31417-help-please.html)

Kerem May 6, 2003 14:26

HELP Please!
 
Hi, I have only one week to finish my graduation project and I need some real help.

I'm trying to analyse the flow around the ROBIN body and compute it's Cd value. I couldn't create a boundry layer for the faces of the ROBIN body because the faces of the body are too complex. So I tried some grid adaptation in Fluent.

[First of all I have to say I compare my results with an experimental study that gives Cd=0.00300 as result. The non-dimensionalization was done by Cd= D/(rho*U^2*R^2) (R is half of the lenght of the ROBIN body)]

1. I did my analyse without adaptation. The mesh on the body face consists of 20000 elements. And I found Cd=0.00369 (but the y+ values on the face are not good. about 150-350 maybe)

2. I did some y+ adaptation and reduce the y+ values to values about 30-80. But with this finer grid I found Cd=0.00400 (That was a surprise!)

3. Then I tried some boundry adaptation for the face increasing the mesh size of the face to 45000 elements and I found Cd=0.00410

4. At last I tried som coarse mesh on face with 5000 face elements and did my analyse without adaptation. I found Cd=0.00375 (again a better result- That was a surprise too)

[I used Coupled-implicit, k-epsilon RNG with non-equilibrium wall functions,discretization scheme all second-order upwind]

Now, I don't understand why I'm away from the right solution when I do the right things (adaptation, fine grid). And I know that too fine grid is not good for the wall function wall treatment and I can say that my grid is not too much fine.

That's all. Anybody with answers to my problem? Please help me!


Kerem May 6, 2003 14:49

Re: HELP Please!
 
Sorry I forgot to say that I used an unstructured volume mesh with tetrahedral elements.

Ahmed May 6, 2003 19:30

Re: HELP Please!
 
Kerem, What was the convegence criterion?.. I suggest you to let the energy and continuity residuals fall below 1e-5 then compare you above solutions.. Good luck

Alex Munoz May 6, 2003 22:02

Re: HELP Please!
 
Hello

I have seen similar behavior in other simmulation performed over cylinder inmersed in a flow. Therefore, Your results are not a surprise for me!

Let me see, If I found some papers for you! I guess the autor is Rodi, The well know German Rodi that has being involved with the development of most of the turbulent models.

Anyway, Search for these papers, They will help you to preper your defence

Large -eddy simmulations of the Flow past Bluff Bodies State of the art Wolfgang Rodi

Oscillating Turbulent Flow with or without a current about a circular cylinder T Sarpkaya et al.

Ok I found it

Search for this key paper

Computation of unsteady viscous flows around bodies using the k-e turbulence model on unstructured dynamic grids. Bruno Koobus, Charbel Farhat and Hai Tran. Actually I have the electronic copy I hope you have enough space in your e-mail box.

Best regards

Alex Munoz


All times are GMT -4. The time now is 18:36.