problems with electro-thermal simulation

Numerical methods and mathematical models of Elmer
Post Reply
lazyTom
Posts: 56
Joined: 05 Jun 2013, 01:32
Antispam: Yes

problems with electro-thermal simulation

Post by lazyTom »

Hi all,

I'm trying to setup a steady state simulation to understand the raise in temperature of a trace on a pcb.

The starting model is very simple, just a small PCB with a straight trace (20mm long). All is immersed in air.

Dimensions are ok, I've checked this by comparing the analytical solution of trace resistance with that of stat current solver.

The problem is that I can't figure out why I get an increase of temperature of 156 deg C in the trace with only 0.5W of Joule losses. Not good...

Joule losses are correct (again matching with theoretical values) so the only place where the problem might be is the heat solver. I've followed the test cases and I've already fixed several bugs that I've made building the sif file but I reached "saturation" and can't see any further :).

Perhaps I'm losing something with thermal BC? Still it doesn't explain a raise in temperature so high

Any pious soul available to have a look at the problem?

I attach an archive with the following files:
.sif
.step
mesh
ELMERSOLVER_STARTINFO

Thanks,

Marco
Attachments
Archivio.zip
(215.13 KiB) Downloaded 237 times
lazyTom
Posts: 56
Joined: 05 Jun 2013, 01:32
Antispam: Yes

Re: problems with electro-thermal simulation

Post by lazyTom »

Hummm... I'm starting to convince myself that theres nothing wrong with the simulation.

In fact, I'm not accounting for any sort convection, heat conductivity of air (0.025) is miserable and if you start to increase it up to that of copper you will see that copper trace temperature (when heat conductivity or air = heat conductivity of copper) will be practically equal to that set on BC.

I think I'll have to change the way I'm approaching the problem.

I will remove the air and use Heat flux boundary conditions to PCB+trace surface and see how it goes.

This brings me to the following question:

Elmer Models Manual says that in order to use Heat Flux boundary condition I need to set

Heat Flux BC = True

but if I look at test cases, this is not always the case. For example in TravellingHeater Heat flux transfer coefficient and External Temperature are set but Heat Flux BC is not.

Any clarification is much appreciated.

Thanks,

Marco
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: problems with electro-thermal simulation

Post by raback »

Hi

"Heat Flux BC" is more or less obsolite since heat flux will be considered if any related keyword is given. I think that currently it has an effect only if given value "False".

-Peter
lazyTom
Posts: 56
Joined: 05 Jun 2013, 01:32
Antispam: Yes

Re: problems with electro-thermal simulation

Post by lazyTom »

Thanks for clarifying!

Marco
mzenker
Posts: 1999
Joined: 07 Dec 2009, 11:49
Location: Germany

Re: problems with electro-thermal simulation

Post by mzenker »

Hi,

did you solve your problem in the meantime?
Otherwise I had a short curious look at your sif and found that you do not do "coordinate scaling = 0.001" in the Simulation section. So if your geometry (which I did not check) is defined in mm, you would have to add the above command since Elmer assumes m by default.

HTH,
Matthias
Post Reply