coupling thermal and electrostatic solvers indirectly

[aminer2k]

Hello,
I would like any advice anyone would have about how to build an analysis that has the thermal and electrostatic solvers indirectly coupled. This is one way to include thermoelectric effects when they are not included in any solver. For example:
1. The solver would iterate the thermal solution (once or several times) and output some line results or surface integrals (using savescalars)
2. Based on these thermal results, a boundary condition for the electrostatic solver would be set (for example, V at a boundary is set based on a function of the thermal results)
3. The electrostatic solution iterates once or several times.
4. The solver goes back to step 1, solving the thermal problem that now includes the joule heating from the electrostatic problem, and repeats this till the problem converges.

I did this in the past using CFDRC-ACE (ESI now), using a user subroutine capability in the software written in fortran.

Is there a way to do this in Elmer? To possible options:
1. do it some way all inside the sif file.
2. write some python to modify the bc in the sif file, and run the loop.

Thanks
Andy

[raback]

Hi

This should be pretty basic stuff for Elmer. You could start by looking at $ELMER_HOME/tests/current_heat_control. This should be doable with the GUI also, little bit depending what kind of feedback you have though.

-Peter

[salad]

Hi Peter,

I am also interested in this electrical-thermal coupling work. I looked the example, current_heat_control, and believe it is a very good case to start with.

A question is about the Potential boundary. I set Potential = 1 at a boundary and 0 at another. Does it mean 1 volt and earth? but in the results, viewed by ParaView, the boundary set with Potential 1 has potential value ~761. I don't understand why. The 0 potential boundary still has 0.

Another question is, how can I plot current distribution and ohimc heat distribution across the domain?

Thanks a lot for all your help.

Best regards,
http://code-saturne.blogspot.com/