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.

Inducing torque at the ends of a deformed, imported 3D beam geometry

Nicholas Goldring Certified Consultant

Please login with a confirmed email address before reporting spam

Hello everyone, I was hoping for some ideas on how to induce a bending moment or torque on a deformed geometry that I have imported from a previous study. The geometry is a relatively large (~1m long) rectangular beam that is very slightly deformed in each direction (μm's). I am currently sticking an arm on either side of the beam and applying a force to the bottoms to induce a moment (please see attached image).

This is unfortunately quite sloppy since the beam is deformed - I have lots of overlap and it's difficult to clean up for meshing. I am currently just reducing repair tolerance and using virtual operations to clean it up but I'm not sure exactly which virtual operations are needed. Typical error messages when trying to mesh this are "overlapping faces" and "failed to insert point".

If anyone has any ideas on how to do induce the moments C_1 and C_2 more gracefully or implement the arms without overlap or any tips on which virtual operations could best be employed here, I would be extremely grateful. Thank you very much!


11 Replies Last Post Jan 31, 2017, 4:58 p.m. EST
Jeff Hiller COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 8 years ago Jan 27, 2017, 3:10 p.m. EST
Hi Nicholas,
Have you considered applying a spatially-dependent load on each end of your beam?
A load that's positive at the top, same magnitude but negative at the bottom, and varies linearly in between will apply a torque without applying a net force. From a Saint-Venant's principle perspective it is the same as a concentrated torque.
Best,
Jeff
Hi Nicholas, Have you considered applying a spatially-dependent load on each end of your beam? A load that's positive at the top, same magnitude but negative at the bottom, and varies linearly in between will apply a torque without applying a net force. From a Saint-Venant's principle perspective it is the same as a concentrated torque. Best, Jeff

Nicholas Goldring Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 8 years ago Jan 27, 2017, 3:14 p.m. EST

Hi Nicholas,
Have you considered applying a spatially-dependent load on each end of your beam?
A load that's positive at the top, same magnitude but negative at the bottom, and varies linearly in between will apply a torque without applying a net force. From a Saint-Venant's principle perspective it is the same as a concentrated torque.
Best,
Jeff


I did not consider this, it sounds great! Do you mean applying the loads on the the upper and lower edges of the beam in the longitudinal direction?

Also, what sort of boundary conditions could I use for this? In some cases this beam will be deformed asymmetrically and I will want to apply moments/torques asymmetrically. The advantage of using the arms was that I could apply my constraints mostly to them without having to place any non-physical constraints on the beam. Right now the only constraint on the beam is an x,y-fixed-point on the bottom-center so it is free to move vertically which will be necessary in the asymmetric cases. I can't find a way to hold the beam in three dimensions without interfering with this asymmetric bending somehow.

Thanks a ton for your reply.

[QUOTE] Hi Nicholas, Have you considered applying a spatially-dependent load on each end of your beam? A load that's positive at the top, same magnitude but negative at the bottom, and varies linearly in between will apply a torque without applying a net force. From a Saint-Venant's principle perspective it is the same as a concentrated torque. Best, Jeff [/QUOTE] I did not consider this, it sounds great! Do you mean applying the loads on the the upper and lower edges of the beam in the longitudinal direction? Also, what sort of boundary conditions could I use for this? In some cases this beam will be deformed asymmetrically and I will want to apply moments/torques asymmetrically. The advantage of using the arms was that I could apply my constraints mostly to them without having to place any non-physical constraints on the beam. Right now the only constraint on the beam is an x,y-fixed-point on the bottom-center so it is free to move vertically which will be necessary in the asymmetric cases. I can't find a way to hold the beam in three dimensions without interfering with this asymmetric bending somehow. Thanks a ton for your reply.

Jeff Hiller COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 8 years ago Jan 27, 2017, 5:11 p.m. EST
See the attached file.
Note that imposing no displacement at a single point is not enough constraints to remove all three rigid body motions in 2D: the rotation around the fixed point remains unchecked. Ultimately, how you constrain the structure needs to be decided based on the physical reality you're modeling.
Best,
Jeff
See the attached file. Note that imposing no displacement at a single point is not enough constraints to remove all three rigid body motions in 2D: the rotation around the fixed point remains unchecked. Ultimately, how you constrain the structure needs to be decided based on the physical reality you're modeling. Best, Jeff


Nicholas Goldring Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 8 years ago Jan 27, 2017, 5:46 p.m. EST

See the attached file.
Note that imposing no displacement at a single point is not enough constraints to remove all three rigid body motions in 2D: the rotation around the fixed point remains unchecked. Ultimately, how you constrain the structure needs to be decided based on the physical reality you're modeling.
Best,
Jeff


Awesome, I'll check this out first thing Tuesday. Thanks so much for putting that together for me. I do understand that the model needs to be constrained in three dimensions - it's just the case that I don't want to impose a vertical constraint on the beam itself because the study is based on vertical bending in order to correct vertical deformation. If the deformations were always symmetric, then I could feasibly constrain some center part of the beam in all 3 dimensions. At any rate, I will look in to it further next week; Have a great weekend!
[QUOTE] See the attached file. Note that imposing no displacement at a single point is not enough constraints to remove all three rigid body motions in 2D: the rotation around the fixed point remains unchecked. Ultimately, how you constrain the structure needs to be decided based on the physical reality you're modeling. Best, Jeff [/QUOTE] Awesome, I'll check this out first thing Tuesday. Thanks so much for putting that together for me. I do understand that the model needs to be constrained in three dimensions - it's just the case that I don't want to impose a vertical constraint on the beam itself because the study is based on vertical bending in order to correct vertical deformation. If the deformations were always symmetric, then I could feasibly constrain some center part of the beam in all 3 dimensions. At any rate, I will look in to it further next week; Have a great weekend!

Nicholas Goldring Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 8 years ago Jan 31, 2017, 9:54 a.m. EST

See the attached file.
Note that imposing no displacement at a single point is not enough constraints to remove all three rigid body motions in 2D: the rotation around the fixed point remains unchecked. Ultimately, how you constrain the structure needs to be decided based on the physical reality you're modeling.
Best,
Jeff


Hi Jeff, would it be possible to get a version of that file that I can open up with version 5.2? Thank you!
[QUOTE] See the attached file. Note that imposing no displacement at a single point is not enough constraints to remove all three rigid body motions in 2D: the rotation around the fixed point remains unchecked. Ultimately, how you constrain the structure needs to be decided based on the physical reality you're modeling. Best, Jeff [/QUOTE] Hi Jeff, would it be possible to get a version of that file that I can open up with version 5.2? Thank you!

Jeff Hiller COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 8 years ago Jan 31, 2017, 2:34 p.m. EST
Here you go!
I still would strongly recommend you upgrade to 5.2a if you can. The name does not necessarily suggest it but there are many valuable additional features in 5.2a compared to 5.2, see www.comsol.com/release/5.2a .
Best,
Jeff
Here you go! I still would strongly recommend you upgrade to 5.2a if you can. The name does not necessarily suggest it but there are many valuable additional features in 5.2a compared to 5.2, see https://www.comsol.com/release/5.2a . Best, Jeff


Nicholas Goldring Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 8 years ago Jan 31, 2017, 2:35 p.m. EST

Here you go!
I still would strongly recommend you upgrade to 5.2a if you can. The name does not necessarily suggest it but there are many valuable additional features in 5.2a compared to 5.2, see www.comsol.com/release/5.2a .
Best,
Jeff

Great, I really appreciate it!
[QUOTE] Here you go! I still would strongly recommend you upgrade to 5.2a if you can. The name does not necessarily suggest it but there are many valuable additional features in 5.2a compared to 5.2, see https://www.comsol.com/release/5.2a . Best, Jeff [/QUOTE] Great, I really appreciate it!

Nicholas Goldring Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 8 years ago Jan 31, 2017, 4:18 p.m. EST
Hmm.. This looks great, but I'm having difficulty applying it to a 3D beam that is quite large.. Any further suggestions? Thank you.
Hmm.. This looks great, but I'm having difficulty applying it to a 3D beam that is quite large.. Any further suggestions? Thank you.

Jeff Hiller COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 8 years ago Jan 31, 2017, 4:22 p.m. EST
The same approach would work in 3D as well, and regardless of the beam size.
I invite you to contact our tech support team if we can be of further help.
Jeff
The same approach would work in 3D as well, and regardless of the beam size. I invite you to contact our tech support team if we can be of further help. Jeff

Nicholas Goldring Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 8 years ago Jan 31, 2017, 4:37 p.m. EST
Thanks, it just doesn't seem to give a real bending moment, the profile is not spherical, it's more just being pulled up at the ends like the image attached.

Thanks, it just doesn't seem to give a real bending moment, the profile is not spherical, it's more just being pulled up at the ends like the image attached.


Jeff Hiller COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 8 years ago Jan 31, 2017, 4:58 p.m. EST
Due to linearization (i.e. the curvature being approximated by the second derivative), you're not expecting a spherical profile but rather a parabolic one. See e.g. An Introduction to the Mechanics of Solids, second edition, by Crandall, Dahl and Lardner, chapter 8.
Jeff
Due to linearization (i.e. the curvature being approximated by the second derivative), you're not expecting a spherical profile but rather a parabolic one. See e.g. An Introduction to the Mechanics of Solids, second edition, by Crandall, Dahl and Lardner, chapter 8. Jeff

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.