CFD Online Discussion Forums

CFD Online Discussion Forums (
-   FLUENT (
-   -   Conjugate heat transfer. (

A7A April 12, 2012 14:47

heat transfer problem.

I am modeling a 2d ribbed channel conjugate heat transfer problem, and i dont know why Nusselt number isnt converging, it continuously decreasing and i dont get the right answer, is it because of the mesh quality or what?

The mesh quality:
Maximum cell squish = 5.56618e-01
Maximum cell skewness = 7.85956e-01
Maximum aspect ratio = 8.35512e+00

Thank you.

Zigainer April 12, 2012 15:56

Can you add a picture of your mesh.
What are your boundary conditions? Which turbulence model?

A7A April 12, 2012 16:06

1 Attachment(s)
Hi Zigainer.
I am using SKW turbulence model, and periodic boundary condition at a flow rate 0.0963 kg/s, and heat flux 2500 W/m2.

Zigainer April 12, 2012 16:25

Have you checked your y+? I guess you want to resolve the boundary layer and not using wall function.
Have you thought about using a "prims" layer for the boundary ....
Could you post a picture of your near wall mesh? Maybe you have to less cells within the boundary layer.

edit: and btw can you change your title, because this is not a conjugate heat transfer problem ;-) (you only simulate the fluid and not fluid&solid)

A7A April 12, 2012 16:44

Thank you for the correction, and about checking y+ I've tried to make y+ adaption in fluent and it told me that there is 0 cells that have y+ bigger than 1, Finally could you tell me more about using a "prims" layer for the boundary.

Zigainer April 12, 2012 16:50

When you do not want to use wall function it is not only important that y+ is less than 1, fluent also requires about 10 cells within the boundary layer. the y+ you get is only of the first cell.

For prism layer:

If you make a grid like this, you do not ne adaption and you have good control of the boundary layer (y+ and numbers of cells).

All times are GMT -4. The time now is 21:10.