Page 1 of 1

Thermal coupling of ttwo boundaries possible?

Posted: 09 Nov 2012, 09:02
by carstenp
Hello everyone,

is it possible to couple two boundaries in elmer together? The boundaries are located at different sides on the mesh and have a slightly different shape. What I want to do is compute the average temperature on boundary one and use this value as external temperature for boundary two and vice verse compute the average temperature on boundary two and use it as external temperature for boundary one. Since the boundaries differ slightly I cant use the head gap model.

So is it possible to model this connection in elmer?

Regards,

carsten

Re: Thermal coupling of ttwo boundaries possible?

Posted: 09 Nov 2012, 10:53
by mzenker
Hi,

I would guess that a user defined function is the best way to do that. But there might be a more elegant way (periodic BCs ?).

HTH,

Matthias

Re: Thermal coupling of ttwo boundaries possible?

Posted: 09 Nov 2012, 11:56
by carstenp
Hi Matthias,

the problem is that the boundaries have a slightly different shape and a different mesh on them so I can't use periodic boundary conditions.

carsten

Re: Thermal coupling of ttwo boundaries possible?

Posted: 09 Nov 2012, 12:37
by mzenker
Hi Carsten,

then I would make a UDF.

Matthias

Re: Thermal coupling of ttwo boundaries possible?

Posted: 09 Nov 2012, 13:38
by carstenp
Ok, so there is no way to use a combination of MATC and SaveScalars?

Re: Thermal coupling of ttwo boundaries possible?

Posted: 09 Nov 2012, 15:27
by mzenker
In principle yes. I just proposed the UDF since you want to do some averaging over your surfaces, and I don't see how this is done with MATC (but I may miss something). But this might also be possible with savescalars (check in the Models Manual), so MATC+savescalars might work here as well. Generally spoken MATC tends to be the slowest variant, if computation time is an issue.

HTH,

Matthias

Re: Thermal coupling of ttwo boundaries possible?

Posted: 09 Nov 2012, 15:40
by carstenp
The averaging would be a nice feature but is not mandatory. It would sufficient to take the temperature at one point on each boundary and use it as external temperature for the other. Computation speed is not an issue, rather to have a really simple solution.

Btw. I had a look into the solver manual to see how I can access nodal values on different boundaries but as I understand it, I can only access data from boundary N when applying a UDF to boundary N,

Re: Thermal coupling of ttwo boundaries possible?

Posted: 09 Nov 2012, 15:56
by mzenker
Hi,

if you really need nodal values, I think you need to use a UDF.
If not, you can use the operators (Min, Max, ...) in SaveScalars (see Models Manual).

HTH,

Matthias