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.

Shell for partially overlapped solid

Please login with a confirmed email address before reporting spam

I want to use shell elements to replace the solid elements in which two thin solids are partially overlapped. Eigenanalysis is conducted. The two partially overlapped (connected) shells seem to be no connected indeed. What should I do to correct the error? Thank you.

The COMSOL 6.1 file is attached.



5 Replies Last Post 09.11.2022, 01:25 GMT-5
Mark Cops COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 2 years ago 08.11.2022, 09:30 GMT-5
Updated: 2 years ago 08.11.2022, 10:45 GMT-5

Dear Panson,

In your model, it looks like you are specifying two different shell thickness types at two different parts of the plate. In the model, I do not see a condition specifying any continuity in displacement where the shells would overlap. One approach would be to instead select three different boundaries for the shell interface (in your model # 4, 8, and 12). Now specify three different thickness conditions for the three separate sections, as shown in the screenshot. In the middle section, the thickness is the sum of the other sections. Each section also needs an appropriate offset defined. Since boundaries 4,8, and 12 in the Shell physics share a common edge, the continuity condition is met.

In your model, a good benchmark would be to first use Solid Mechanics and solve for the eigenfrequencies and compare to the Shell solution.

Best,

Mark

Dear Panson, In your model, it looks like you are specifying two different shell thickness types at two different parts of the plate. In the model, I do not see a condition specifying any continuity in displacement where the shells would overlap. One approach would be to instead select three different boundaries for the shell interface (in your model # 4, 8, and 12). Now specify three different thickness conditions for the three separate sections, as shown in the screenshot. In the middle section, the thickness is the sum of the other sections. Each section also needs an appropriate offset defined. Since boundaries 4,8, and 12 in the Shell physics share a common edge, the continuity condition is met. In your model, a good benchmark would be to first use Solid Mechanics and solve for the eigenfrequencies and compare to the Shell solution. Best, Mark


Henrik Sönnerlind COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 2 years ago 08.11.2022, 10:32 GMT-5
Updated: 2 years ago 08.11.2022, 11:03 GMT-5

There is an easier way:

Add a Boundary to Boundary connection from Connections at the boundary level in the Shell interface.

However, when working with shells and offsets, you have to be careful about what is top and bottom. Add a Shell Geometry plot from the Add Predefined Plot window to verify.

-------------------
Henrik Sönnerlind
COMSOL
There is an easier way: Add a *Boundary to Boundary* connection from *Connections* at the boundary level in the *Shell* interface. However, when working with shells and offsets, you have to be careful about what is top and bottom. Add a *Shell Geometry* plot from the *Add Predefined Plot* window to verify.

Please login with a confirmed email address before reporting spam

Posted: 2 years ago 09.11.2022, 00:13 GMT-5
Updated: 2 years ago 09.11.2022, 00:18 GMT-5

Dear Panson,

In your model, it looks like you are specifying two different shell thickness types at two different parts of the plate. In the model, I do not see a condition specifying any continuity in displacement where the shells would overlap. One approach would be to instead select three different boundaries for the shell interface (in your model # 4, 8, and 12). Now specify three different thickness conditions for the three separate sections, as shown in the screenshot. In the middle section, the thickness is the sum of the other sections. Each section also needs an appropriate offset defined. Since boundaries 4,8, and 12 in the Shell physics share a common edge, the continuity condition is met.

In your model, a good benchmark would be to first use Solid Mechanics and solve for the eigenfrequencies and compare to the Shell solution.

Best,

Mark

Dear Mark,

Thank you for your response. I fogot to mention and define in the model that the two different plates are with different materials. If I follow your suggestion, is the middle section a layered shell like that in the example "failure predication in a layered shell"? I should then follow the steps in the example for the middle section. Is it correct?

Thank you.

Panson

>Dear Panson, > >In your model, it looks like you are specifying two different shell thickness types at two different parts of the plate. In the model, I do not see a condition specifying any continuity in displacement where the shells would overlap. One approach would be to instead select three different boundaries for the shell interface (in your model # 4, 8, and 12). Now specify three different thickness conditions for the three separate sections, as shown in the screenshot. In the middle section, the thickness is the sum of the other sections. Each section also needs an appropriate offset defined. Since boundaries 4,8, and 12 in the Shell physics share a common edge, the continuity condition is met. > >In your model, a good benchmark would be to first use Solid Mechanics and solve for the eigenfrequencies and compare to the Shell solution. > >Best, > >Mark Dear Mark, Thank you for your response. I fogot to mention and define in the model that the two different plates are with different materials. If I follow your suggestion, is the middle section a layered shell like that in the example "failure predication in a layered shell"? I should then follow the steps in the example for the middle section. Is it correct? Thank you. Panson

Please login with a confirmed email address before reporting spam

Posted: 2 years ago 09.11.2022, 01:08 GMT-5
Updated: 2 years ago 09.11.2022, 01:09 GMT-5

There is an easier way:

Add a Boundary to Boundary connection from Connections at the boundary level in the Shell interface.

However, when working with shells and offsets, you have to be careful about what is top and bottom. Add a Shell Geometry plot from the Add Predefined Plot window to verify.

Dear Henrik,

Thank you!

I have the Shell Geometry plot as the attached image file. I think the setting for thickness and offset are correct.

>There is an easier way: > >Add a *Boundary to Boundary* connection from *Connections* at the boundary level in the *Shell* interface. > >However, when working with shells and offsets, you have to be careful about what is top and bottom. Add a *Shell Geometry* plot from the *Add Predefined Plot* window to verify. Dear Henrik, Thank you! I have the Shell Geometry plot as the attached image file. I think the setting for thickness and offset are correct.


Please login with a confirmed email address before reporting spam

Posted: 2 years ago 09.11.2022, 01:25 GMT-5
Updated: 2 years ago 09.11.2022, 01:25 GMT-5

I use boundary to boundary connection suggested by Henrik. Mesh density for solid/shell model is shown in the attached picture. Eigenfrequencies for both models are in good agreement. Is the increasing discrepancy against frequency an expected outcome for using shell? The bending curvature is steeper for higher eigenfrequency for which shell approximation is degrading. Is it a proper interpation?

I use boundary to boundary connection suggested by Henrik. Mesh density for solid/shell model is shown in the attached picture. Eigenfrequencies for both models are in good agreement. Is the increasing discrepancy against frequency an expected outcome for using shell? The bending curvature is steeper for higher eigenfrequency for which shell approximation is degrading. Is it a proper interpation?

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.