CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   heat flux on fluid-solid interface does not match (

Dai October 27, 2006 02:41

heat flux on fluid-solid interface does not match
Hi, everybody,

I have a fluid domain with solid enclosures touching the boundary. Sources of heat are located in fluid. In 2D everything works, but in 3D heat flux does not match on solid and fluid interface of enclosures.

I have made a mesh in Workbench by meshing one single part, which consist of 3 bodies. (2 enclosures are add frozen extrudes). Visually it seems that mesh nodes fit on the interface...

Where could be the problem? BC on interface is interface flux, I have tries both CGI and 1:1...


opaque October 27, 2006 07:49

Re: heat flux on fluid-solid interface does not ma
Dear Dai,

How are you comparing if the heat fluxes match? Using CFX-Post, or by looking at the ANSYS CFX solver output file..


Robin October 27, 2006 08:54

Re: heat flux on fluid-solid interface does not ma
Heat flux is strictly conserved at an interface. This is how the GGI actually works, so the problem must be how you are interpreting the results. Are you comparing visually or quantitatively. Provide the details and we can probably find the problem.


Dai October 28, 2006 08:45

Re: heat flux on fluid-solid interface does not ma
I have concluded that it could be because of not enough elements in the direction of heat flux... (I have a solid planar enclosure in the fluid flow and heat flux must go through it perpendiculary to the plane - at the moment CGI gives not continuos T isolines (they are separate set for solid and fluid), but in 1:1 method isolines are continous, but the solid enclosure is heated up and heat does not go away...

I have made the mesh in Workbench and till now I have not found how I can mesh this planar enclosure separately with elements which are appropriate to object with one dimension << other dimensions...So at the moment that is the problem I want to find out - how can I set different types of elements in Workbench to different bodies?

Robin October 30, 2006 08:38

Re: heat flux on fluid-solid interface does not ma
Hi Dai,

First of all, it doesn't sound like you have a 1:1 connection between nodes, which is why the heat doesn't go away when you force a 1:1 interface. Forcing 1:1 simply prevents the solver from a using a GGI at the interface, it does not make it a conformal mesh.

Secondly, you can't judge this based on a plot. There will be discontinuities at your interface because of changes in mesh size and changes in the materials. Discontinuities can also arise in your plot if you are looking at hybrid values, try conservative values instead.

As I said before, the GGI strictly conserves heat flux at the interface, so the problem is with your interpretation of the results. The best way to know what you have done wrong is if you tell us what you are plotting in detail. I.e. what variable are you trying to compare, how are you calculating heat flux, etc.

Regards, Robin

All times are GMT -4. The time now is 13:01.