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 The information provided may be out of date.

Discussion Locked This discussion was locked by a forum moderator.

How to deal with the nonconvergence at interface in solid mechanics

Please login with a confirmed email address before reporting spam

Hi,

Recently I encountered a problem. When I used solid mechanics module to simulate the thermal expansion deformation of a particle-matrix system(a sphere particle included in a matrix, axi-symmetric model). I cannot converge at some time. As fig. t430 and t440 shown, when time is 430s, the result is right while when time is 440s, the calculation cannot converge. However, if I only simulate the expansion of this particle, it can work well. So I want to make sure if it is the interface causes the problem. Is there any ways to deal with it? Thanks!

Best,
Rong


8 Replies Last Post Jun 1, 2015, 6:15 p.m. EDT

Please login with a confirmed email address before reporting spam

Posted: 9 years ago May 28, 2015, 4:53 p.m. EDT
Hi,

I also found when I even added a homogeneous temperature field on the center part, it cannot converge at some time. I think it is a very simple problem with large deformation, and in ABAQUS it can work well. But I don't know why I cannot handle with it in COMSOL. Attached is the pressure at different steps. Thanks!

Best,
Rong
Hi, I also found when I even added a homogeneous temperature field on the center part, it cannot converge at some time. I think it is a very simple problem with large deformation, and in ABAQUS it can work well. But I don't know why I cannot handle with it in COMSOL. Attached is the pressure at different steps. Thanks! Best, Rong


Please login with a confirmed email address before reporting spam

Posted: 9 years ago May 28, 2015, 5:14 p.m. EDT
Attached is the result from ABAQUS.
Attached is the result from ABAQUS.


Please login with a confirmed email address before reporting spam

Posted: 9 years ago May 28, 2015, 6:27 p.m. EDT
Hi,

I uploaded this COMSOL model. Can anyone help me check this model? It is a very simple model. I just added a thermal stain and want to get the stress distribution. Thanks!

Best,
Rong
Hi, I uploaded this COMSOL model. Can anyone help me check this model? It is a very simple model. I just added a thermal stain and want to get the stress distribution. Thanks! Best, Rong


Jeff Hiller COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 9 years ago May 29, 2015, 8:20 a.m. EDT
Hello Rong,
For one, your model is not set up as axisymmetric.
Another problem is that the coefficient of thermal expansion for material 1 is set to a wild value (1[1/K]! ).
I hope this helps.
Best regards,
Jeff
Hello Rong, For one, your model is not set up as axisymmetric. Another problem is that the coefficient of thermal expansion for material 1 is set to a wild value (1[1/K]! ). I hope this helps. Best regards, Jeff

Please login with a confirmed email address before reporting spam

Posted: 9 years ago May 29, 2015, 1:50 p.m. EDT
Hi Jeff,

Thanks so much! This model is just a simplified model. Even I build an axis-symmetric model (attached is that axis-symmetric model). It cannot converge yet. So I think it is not because of the axis-symmetric frame. Is it because the difference of Young's Module between two materials is two large?
According to the thermal expansion coefficient, finally I need a 50% thermal strain, so for simplification I set CTE equals to 1 and the temperature equals 50 0.5. Of cause I can define CTE as a smaller value while the temperature gradient would increase. So I think the result is the same. Thanks!

Best,
Rong
Hi Jeff, Thanks so much! This model is just a simplified model. Even I build an axis-symmetric model (attached is that axis-symmetric model). It cannot converge yet. So I think it is not because of the axis-symmetric frame. Is it because the difference of Young's Module between two materials is two large? According to the thermal expansion coefficient, finally I need a 50% thermal strain, so for simplification I set CTE equals to 1 and the temperature equals 50 0.5. Of cause I can define CTE as a smaller value while the temperature gradient would increase. So I think the result is the same. Thanks! Best, Rong


Henrik Sönnerlind COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 9 years ago May 31, 2015, 5:19 p.m. EDT
Hi,

If you will have 50% strain, linear elasticity is not a suitable model. It will break down at large compressive strains.

There are at least two issues here:

1. Such a large thermal strain requires a multiplicative decomposition. This is used in the hyperelastic materials.

2. Interpretation of linear elasticity is different for different choices of stress and strain measures. See also www.comsol.com/blogs/why-all-these-stresses-and-strains . Two different softwares having different formulations will give vastly different results at 50% strain and linear elasticity. This is what you experienced in your first post.

Regards,
Henrik
Hi, If you will have 50% strain, linear elasticity is not a suitable model. It will break down at large compressive strains. There are at least two issues here: 1. Such a large thermal strain requires a multiplicative decomposition. This is used in the hyperelastic materials. 2. Interpretation of linear elasticity is different for different choices of stress and strain measures. See also http://www.comsol.com/blogs/why-all-these-stresses-and-strains . Two different softwares having different formulations will give vastly different results at 50% strain and linear elasticity. This is what you experienced in your first post. Regards, Henrik

Please login with a confirmed email address before reporting spam

Posted: 9 years ago Jun 1, 2015, 6:10 p.m. EDT

Hi Henrik,

Thanks for your reply. Indeed, I have open the large deformation option including large plastic strain, so I think in my model the geometric nonlinear problem has been considered.
You said a large thermal strain requires a multiplicative decomposition. I think what you mean is the elastic deformation induced by thermal strain requires a multiplicative decomposition. However, is it right the multiplicative decomposition has been done when I used large plastic strain model? Now the elastic strain is the green-Lagrangian strain. But the model still cannot work. Do you have some advice for dealing with such a large thermal strain? Thank!

Best,
Rong
Hi Henrik, Thanks for your reply. Indeed, I have open the large deformation option including large plastic strain, so I think in my model the geometric nonlinear problem has been considered. You said a large thermal strain requires a multiplicative decomposition. I think what you mean is the elastic deformation induced by thermal strain requires a multiplicative decomposition. However, is it right the multiplicative decomposition has been done when I used large plastic strain model? Now the elastic strain is the green-Lagrangian strain. But the model still cannot work. Do you have some advice for dealing with such a large thermal strain? Thank! Best, Rong

Please login with a confirmed email address before reporting spam

Posted: 9 years ago Jun 1, 2015, 6:15 p.m. EDT

Hi Henrik,

I suddenly realized that only the center part of my model have considered the plasticity as well as multiplicative decomposition you mentioned while outer shell is pure elasticity. So maybe it is the point in your reply! I will modify my model soon to see if it is right. Thanks again.

Best,
Rong
Hi Henrik, I suddenly realized that only the center part of my model have considered the plasticity as well as multiplicative decomposition you mentioned while outer shell is pure elasticity. So maybe it is the point in your reply! I will modify my model soon to see if it is right. Thanks again. Best, Rong

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.