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 The information provided may be out of date.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.

von Mises stress too high??

Please login with a confirmed email address before reporting spam

Hi,
I followed the busbar tutorial found in the "Introduction to COMSOL Multiphysics" booklet. The model is also available in the "Model Library.
I got the exactly same solution as shown in the booklet. The temperature of the busbar is from 330K-322K. At such low temperature, the von Mises stress is very high (max=122 MPa), much higher than I expected.
The copper piece, according to the COMSOL calculation results, is yielding!!

I also used solid mechanics in a few of my other models, the von Mises stress I got from COMSOL are all way too high (>>yield strength of materials) than I expected. But in reality, these designs work OK.

Is there anything wrong with my interpretation of the data or understanding of solid mechanics?

Thanks!!

11 Replies Last Post 26 giu 2012, 16:22 GMT-4
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 22 giu 2012, 05:02 GMT-4
Hi

it could be an mesh density issue, if you have stress concentration points. But also what is "too high" <5% 50% ...

So far, over the last years I have never observed significant differences from COMSOL results than from other FEM tools, in my solid models

--
Good luck
Ivar
Hi it could be an mesh density issue, if you have stress concentration points. But also what is "too high"

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 25 giu 2012, 11:14 GMT-4
Hi Friends,

I also need help in the same thing. I am using different software and in my model, Von mises stress is too higher than the input values of uni axial stress values. Sometimes it will take 5 times than my input values. If anybody know this reason, please share with me.

Thanks for your help.
Hi Friends, I also need help in the same thing. I am using different software and in my model, Von mises stress is too higher than the input values of uni axial stress values. Sometimes it will take 5 times than my input values. If anybody know this reason, please share with me. Thanks for your help.

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 25 giu 2012, 14:07 GMT-4
Hi

stress build-up is very model and mesh dependent are you sure you have "exactly" the same model and settings, including mesh and discretization ?

--
Good luck
Ivar
Hi stress build-up is very model and mesh dependent are you sure you have "exactly" the same model and settings, including mesh and discretization ? -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 25 giu 2012, 16:28 GMT-4
Hi,
I tried to refine meshes, but reducing mesh size further cannot give any significant changes in the results. My original mesh size is already fine enough.
The maximum von Mises stress in my model is ridiculously high. If it was the case, the piece I am modeling would brake in real life. But in real life, it never happened.
Hi, I tried to refine meshes, but reducing mesh size further cannot give any significant changes in the results. My original mesh size is already fine enough. The maximum von Mises stress in my model is ridiculously high. If it was the case, the piece I am modeling would brake in real life. But in real life, it never happened.

Nagi Elabbasi Facebook Reality Labs

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 25 giu 2012, 18:12 GMT-4
It is common to get unrealistic high stresses in thermal-structural problems due to improper (over-constrained) structural boundary conditions. For example, if you take a copper bar and fix it at both ends and subject it to a uniform 40oC temperature rise the stresses will exceed yield. In practice the supports usually have a little flexibility that will avoid these extremely high stresses.

Nagi Elabbasi
Veryst Engineering
It is common to get unrealistic high stresses in thermal-structural problems due to improper (over-constrained) structural boundary conditions. For example, if you take a copper bar and fix it at both ends and subject it to a uniform 40oC temperature rise the stresses will exceed yield. In practice the supports usually have a little flexibility that will avoid these extremely high stresses. Nagi Elabbasi Veryst Engineering

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 25 giu 2012, 18:19 GMT-4
Hi Nagi,
Thanks for your help.
In my model, I set no fixed constraint or constrain only one face. But in both cases I still got unrealistically high thermal stress.
Hi Nagi, Thanks for your help. In my model, I set no fixed constraint or constrain only one face. But in both cases I still got unrealistically high thermal stress.

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 26 giu 2012, 01:43 GMT-4
Hi

thanks Nagi for pointing to the thermo elasticity, indeed a Fixed constraint is "fixed", if you have thermal aspects what I do is to give a material expansion on the "fixed interface". The easiest is to use the prescribed displacement node and program in the thermal expansion of the material of the fixed fixation. but you still need to define a "Centre of Expansion" I do that mostly by taking the average position in X,Y,Z of the fixed surface, but it depends on the fixation means, if a few pins are the reference then its the average position of the pin axis one could consider. This gives something like displacement in x := (x-X0)*alpha_fix*(T-Tref), one can also use the local coordinates

I have made a simple 4.3 model with load cases, these are nice ;)

--
Good luck
Ivar
Hi thanks Nagi for pointing to the thermo elasticity, indeed a Fixed constraint is "fixed", if you have thermal aspects what I do is to give a material expansion on the "fixed interface". The easiest is to use the prescribed displacement node and program in the thermal expansion of the material of the fixed fixation. but you still need to define a "Centre of Expansion" I do that mostly by taking the average position in X,Y,Z of the fixed surface, but it depends on the fixation means, if a few pins are the reference then its the average position of the pin axis one could consider. This gives something like displacement in x := (x-X0)*alpha_fix*(T-Tref), one can also use the local coordinates I have made a simple 4.3 model with load cases, these are nice ;) -- Good luck Ivar


Nagi Elabbasi Facebook Reality Labs

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 26 giu 2012, 15:43 GMT-4
Thanks Ivar, that’s a nice example.

Freebody, one other source of high thermal stresses is mismatch in coefficient of thermal expansion between two bonded materials at their interface.
Thanks Ivar, that’s a nice example. Freebody, one other source of high thermal stresses is mismatch in coefficient of thermal expansion between two bonded materials at their interface.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 26 giu 2012, 16:14 GMT-4
Hi Ivar,
Thanks a lot for your example.
Hi Ivar, Thanks a lot for your example.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 26 giu 2012, 16:17 GMT-4
Hi Nagi,
Thanks for your help.
I'll continue investigate the problem I have. I'll post more comments when I find the causes of my problem.
The geometry of my model actually only have one domain (one material) although its shape is complicated and has some sharp edges.
Hi Nagi, Thanks for your help. I'll continue investigate the problem I have. I'll post more comments when I find the causes of my problem. The geometry of my model actually only have one domain (one material) although its shape is complicated and has some sharp edges.

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 26 giu 2012, 16:22 GMT-4
Hi

then just for fun, try to set nu (Poisson) to 0 and solve, you might have shear tress buildup in some regions, these should dissapear with nu=0, but the correct results is still for nu at nominal value ;)

--
Good luck
Ivar
Hi then just for fun, try to set nu (Poisson) to 0 and solve, you might have shear tress buildup in some regions, these should dissapear with nu=0, but the correct results is still for nu at nominal value ;) -- 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.