code_saturne User's Forum

Hi Yvan:

I wish I could have found a more independent help on thermal boundary conditions in code saturne 3.1 but i could not hence writing to you.

I am flexing my hands on free convection problems in complex geometries for which nusselt number correlations do not exist. I am starting with very simple geometries to learn the process.

I have a rectangular geometry with air trapped. Side walls are insulated. Top wall at fixed temperature. Lower wall with unknown temperature but with known heat flux (watt/m2).

I actually want to find pure convection coefficient for this problem using code saturne so i figured I will determine the temperature of lower wall using heat flux. If there is more direct approach to convection coefficient, plz guide.

Overall, my questions are:

1) what is prescribed flux in thermal boundary conditions for scalar temperature? Is it Watt/m2 or something else?
2) Does setting this flux equal to zero suffice for insulated conditions?
3) Does this flux has a sign convention. If i need to simulate heat flux across the lower wall into surroundings, will positive sign suffice?
4) What is exchange coefficient and prescribed value along with it?
5) Can you suggest any built-in and/or systematic approach in determining convection coefficient?
6) In general how can I determine the temperature of lower wall?

Thanks

Hello,

Note that version 3.1 is not supported anymore.

Heat flux is in W/m2. I think a positive flux means outgoing fluw, but I recommend testing both ways to check how the temperature evolves. 0 heat flux means adiabatic condition.

The exchange coefficient my be used if you know how the local exchange behaves. Depending on the case, you may use a prescribed wall temperature. Depending on whether the wall's temperature is maintained externally or changes, you may need to use conjugate heat transfer or not.

But in your case, since the conditions are known (insulated sides, fixed top temperature, fixed bottom heat flux), you probably do not need to specify an exchange coefficient (for the fixed top temperature, you can let the code compute the coefficient automatically based on wall laws, though the quality of this estimation may depend on the mesh quality and turbulence model characteristics).

You can postprocess the temperature of the lower wall easily with recent Code_Saturne versions.

Regards,

Yvan