Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.

Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Thermal stress model - symetry problem

Please login with a confirmed email address before reporting spam

Hi everyone,

I am trying to model the thermal stresses in a thermoelectric module upon cooling from brazing temperature.

In order to save time, I'd like to use symetries in my geometry (I can actually calculate on 1/16th of the total geometry due tu symetries) but therre is a problem during computation and the calcuculation stops.

I checked the boundary conditions and everything looks fine.

Can somebody help me ?

You will find attached the full geometry and the 1/16th geometry.

Thank you in advance


Mathieu


3 Replies Last Post 17.02.2012, 15:16 GMT-5
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 16.02.2012, 17:33 GMT-5
Hi

a few comments:
In the reduces you have dedoubled symmetry conditions, remove or disable half of them
You need to fully define the position for the solid part, in the 1/16 the symmetry BCs fixes the part in space, so the prescribed displacement will stress the device further, perhaps a soft spring would help, or a point load (will give singularities)
The full model is not restrained completely and can "move" this will give solver issues, mot probably
You have, for both models set an initial temperature very high: about 1000K but you keep the Thermal expansion Strain reference temperature at room values, this astonishes me, I would have expected also 1000K (per Thermal linear material node)
You are using a segregated solver, often, if you have enough RAM, the direct solver is more eficient. Try to add a Fully coupled subnode in the solver sequence, this will disble the segregated solver, but you can always enable it to switch back
meshing with only one element per thickness is not ideal, but probably OK for a SS case. You can also use a structured sweep mesh

--
Good luck
Ivar
Hi a few comments: In the reduces you have dedoubled symmetry conditions, remove or disable half of them You need to fully define the position for the solid part, in the 1/16 the symmetry BCs fixes the part in space, so the prescribed displacement will stress the device further, perhaps a soft spring would help, or a point load (will give singularities) The full model is not restrained completely and can "move" this will give solver issues, mot probably You have, for both models set an initial temperature very high: about 1000K but you keep the Thermal expansion Strain reference temperature at room values, this astonishes me, I would have expected also 1000K (per Thermal linear material node) You are using a segregated solver, often, if you have enough RAM, the direct solver is more eficient. Try to add a Fully coupled subnode in the solver sequence, this will disble the segregated solver, but you can always enable it to switch back meshing with only one element per thickness is not ideal, but probably OK for a SS case. You can also use a structured sweep mesh -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 17.02.2012, 04:05 GMT-5
Hi Ivar and thank you for your comments.

I added the prescribed displacement in order to prevent the geometry to rotate but it seems to give rise to problems.

What would you advise me to insert as BC to fix the geometry in space and not constrain it ?

Also I used room temperature for the strain free reference temperature because the geometry drawn is the one at room temperature. Eventually what I'd like to do is to add some plasticity to the commer parts and vary the temperature from room temp to 800°C and then come back to room temp in order to see the deformations in the coppoer and the remaining stresses.

Here is the model corrected with you advice but still a degree of freedom problem.

Thank you
Hi Ivar and thank you for your comments. I added the prescribed displacement in order to prevent the geometry to rotate but it seems to give rise to problems. What would you advise me to insert as BC to fix the geometry in space and not constrain it ? Also I used room temperature for the strain free reference temperature because the geometry drawn is the one at room temperature. Eventually what I'd like to do is to add some plasticity to the commer parts and vary the temperature from room temp to 800°C and then come back to room temp in order to see the deformations in the coppoer and the remaining stresses. Here is the model corrected with you advice but still a degree of freedom problem. Thank you


Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 17.02.2012, 15:16 GMT-5
Hi

one important thing, for your material proerty functions, you should enter the argument as T[1/degC] to have it coverted to unitless and scale of °C, now its flagged orage and probaly T is transferred in Kelvin

The model you uploaded does not have any other solid BC,s apart from the 3 symmetries, that are enogh to fix it in space as is. THen you have applyed a "roller condition" on the top. This is not lear for me, as its very simular to a 4th symmetry condition. Roler means it connaot move along the direction perpendicular to the plane, but freely in any direction "in the plane".

TS solves for T and for u, so you need enough BCs to fix all 1+6 DoFs

--
Good luck
Ivar
Hi one important thing, for your material proerty functions, you should enter the argument as T[1/degC] to have it coverted to unitless and scale of °C, now its flagged orage and probaly T is transferred in Kelvin The model you uploaded does not have any other solid BC,s apart from the 3 symmetries, that are enogh to fix it in space as is. THen you have applyed a "roller condition" on the top. This is not lear for me, as its very simular to a 4th symmetry condition. Roler means it connaot move along the direction perpendicular to the plane, but freely in any direction "in the plane". TS solves for T and for u, so you need enough BCs to fix all 1+6 DoFs -- Good luck Ivar

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.