Menu

Discussion Forum

New Discussion
Filtrer

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.

Why the negative and positive stress?

Posted 23 nov. 2018, 12:43 UTC−5 Heat Transfer & Phase Change, Heat Transfer, Structural Mechanics Version 5.4 6 Replies

Hua Hua
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Hi there,

I'm some confusing on the stress tensor values when conducting the thermal expansion in linear elastic model. Why the y-stress values are oscillating between zero. Is it because of the "smooth"/"resolution" problem in the plot Quality? I checked it but showing positive stress. If it is the tension stress (positive stress), why is the negative displacment? It seems to me that there should be the compression stress (negative stress) because of the negative displancement.

Please check the attached model and figures. As temperature decreasing and cooling (from top side to bottom side), the FREE top end will contract downward. In theory, the top free side shouldn't have vertical thermal stress because of no constrains.

Thank your explanations very much!! Xi

6 Replies Last Post 6 déc. 2018, 10:05 UTC−5
Henrik Sönnerlind COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 6 years ago 26 nov. 2018, 05:04 UTC−5
Updated: 6 years ago 26 nov. 2018, 05:06 UTC−5

Hi,

Avoid high refinement values in plot Quality settings. If values are supposed to be close to zero, you will just get an amplification of small numerical errors.

For thermal stress problems in particular, this will give very strong effects, unless the shape functions for the temperature is one order lower than for the displacements. The reason is that the stress is proportional to the difference between thermal strains (proportional to temperature) and mechanical strains (proportional to derivatives of displacements).

Regards,
Henrik

-------------------
Henrik Sönnerlind
COMSOL
Hi, Avoid high refinement values in plot Quality settings. If values are supposed to be close to zero, you will just get an amplification of small numerical errors. For thermal stress problems in particular, this will give very strong effects, unless the shape functions for the temperature is one order lower than for the displacements. The reason is that the stress is proportional to the difference between thermal strains (proportional to temperature) and mechanical strains (proportional to derivatives of displacements). Regards, Henrik
Hua Hua
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 6 years ago 26 nov. 2018, 15:11 UTC−5
Updated: 6 years ago 21 janv. 2019, 21:25 UTC−5

Hi Henrik,

Thank for your kind replies very much! Ok, I know the oscillating values could be from numerical values. When I choose the "Quality"/"no refinement", the profile looks some smooth but looks like a step-wise curve. Please see the attached figures.

Thanks Xi

Hi Henrik, Thank for your kind replies very much! Ok, I know the oscillating values could be from numerical values. When I choose the "Quality"/"no refinement", the profile looks some smooth but looks like a step-wise curve. Please see the attached figures. Thanks Xi
Henrik Sönnerlind COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 6 years ago 29 nov. 2018, 10:07 UTC−5

Hi Xi,

The stationary plot is just numerical noise (the stress is less than a microPascal).

The step-wise behavior is caused by using a cut line. A line in the geometry will have access to more information, so that proper inter-element averaging can be performed.

As for "shape functions of temperature", I mean the approximation of the temperature field within each element. You can change the order of the shape functions in the Discretization section in the settings for a physics interface. If you add a Thermal Stress multiphysics interface, the shape function order for the temperatures is lowered to Linear by default. When combine your physics interfaces manually, you will typically get Quadratic.

Note that the Y-direction stress due to thermal expansion will be small. The important stresses will occur in the horizontal direction. Which stress measure to use for studying damage will depend in the type of material you are dealing with.

Regards,
Henrik

-------------------
Henrik Sönnerlind
COMSOL
Hi Xi, The stationary plot is just numerical noise (the stress is less than a microPascal). The step-wise behavior is caused by using a cut line. A line in the geometry will have access to more information, so that proper inter-element averaging can be performed. As for "shape functions of temperature", I mean the approximation of the temperature field within each element. You can change the order of the shape functions in the Discretization section in the settings for a physics interface. If you add a Thermal Stress multiphysics interface, the shape function order for the temperatures is lowered to Linear by default. When combine your physics interfaces manually, you will typically get Quadratic. Note that the Y-direction stress due to thermal expansion will be small. The important stresses will occur in the horizontal direction. Which stress measure to use for studying damage will depend in the type of material you are dealing with. Regards, Henrik
Hua Hua
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 6 years ago 30 nov. 2018, 12:04 UTC−5
Updated: 6 years ago 21 janv. 2019, 21:24 UTC−5

Hi Henrik,

Many thanks! I can understand that is numerical noise in the stationary case where the Y-stress should be zero. At this moment, I'm more interested in the vertical stress status. If needed, we can constraint the top boundary.

Best, Xi

Hi Henrik, Many thanks! I can understand that is numerical noise in the stationary case where the Y-stress should be zero. At this moment, I'm more interested in the vertical stress status. If needed, we can constraint the top boundary. Best, Xi
Henrik Sönnerlind COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 6 years ago 6 déc. 2018, 04:17 UTC−5

Hi Xi,

Inter element averaging works better if you have a line in the geometry, rather than a cut line.

Yes, a homogeneous external strain will just move the material, but not cause any stress (as long as the material is free to expand).

Regards,
Henrik

-------------------
Henrik Sönnerlind
COMSOL
Hi Xi, Inter element averaging works better if you have a line in the geometry, rather than a cut line. Yes, a homogeneous external strain will just move the material, but not cause any stress (as long as the material is free to expand). Regards, Henrik
Hua Hua
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 6 years ago 6 déc. 2018, 10:05 UTC−5
Updated: 6 years ago 21 janv. 2019, 21:24 UTC−5

Hi Henrik,

Thanks very much for your kind replies.

Best, Xi

Hi Henrik, Thanks very much for your kind replies. Best, Xi

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.

Suggested Content