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.

Measuring angle of twist

Please login with a confirmed email address before reporting spam

Hi,

I'm building a fairly simple model of a beam that is clamped on one side and subjected to torsion on the other side. Due to the applied load, the beam twists and I want to measure the angle of twist. I tried some probes but I couldn't figure out how to measure the angle of the twisted beam with respect to the untwisted (original) configuration.

I hope my question is clear enough.

Thanks in advance!

3 Replies Last Post 4 oct. 2011, 01:16 UTC−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 3 oct. 2011, 10:50 UTC−4
Hi

there are two ways I know about:

1) for small angles use the antisymmetric curl part of the displacement field or simply in V4, see most theory books or the Timoshenko or ...

Define Variables
Thx = 0.5[rad]*solid.curlUX
Thy = 0.5[rad]*solid.curlUY
Thz = 0.5[rad]*solid.curlUZ

this is fully valid for "small displacements" get wrong above a few degrees. (for the formula of "solid.curlUx" see the "equation tab")

2) another way, also valid for larger displacements is to set a rigid body boundary condition on your extremity where you want to measure the angles, and use the new dependent variables of the angle that COMSOL set ups for you on the Rigid Connector, but there is, today, a serious caveat: the RC also makes the boundary fully RIGID (that's what it's, for OK) but this is not always what is desired.
I have heard that "Soft or average Connectors" are in preparation, not sure when they are coming, perhaps at the next COMSOL conference release, next week, with V4.2a at least I hope so, as well as certainly other goodies :)

--
Good luck
Ivar
Hi there are two ways I know about: 1) for small angles use the antisymmetric curl part of the displacement field or simply in V4, see most theory books or the Timoshenko or ... Define Variables Thx = 0.5[rad]*solid.curlUX Thy = 0.5[rad]*solid.curlUY Thz = 0.5[rad]*solid.curlUZ this is fully valid for "small displacements" get wrong above a few degrees. (for the formula of "solid.curlUx" see the "equation tab") 2) another way, also valid for larger displacements is to set a rigid body boundary condition on your extremity where you want to measure the angles, and use the new dependent variables of the angle that COMSOL set ups for you on the Rigid Connector, but there is, today, a serious caveat: the RC also makes the boundary fully RIGID (that's what it's, for OK) but this is not always what is desired. I have heard that "Soft or average Connectors" are in preparation, not sure when they are coming, perhaps at the next COMSOL conference release, next week, with V4.2a at least I hope so, as well as certainly other goodies :) -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 3 oct. 2011, 16:14 UTC−4
Thanks for your help Ivar, unfortunately this doesn't do it for me.

I'm interesting in large rotations, so your first option will not be correct for me. And the second option cannot be used due to the fact that the boundary becomes fully rigid.

Isn't there a way to define a local coordinate system on the free tip of the beam and measure the rotation of that coordinate system with respect to the global coordinate system? If so, how?

EDIT: Another option would be to prescribe a rotation at the free end and measure the torque. However, I also have no clue how to measure the torque and prescribe the rotation.
Thanks for your help Ivar, unfortunately this doesn't do it for me. I'm interesting in large rotations, so your first option will not be correct for me. And the second option cannot be used due to the fact that the boundary becomes fully rigid. Isn't there a way to define a local coordinate system on the free tip of the beam and measure the rotation of that coordinate system with respect to the global coordinate system? If so, how? EDIT: Another option would be to prescribe a rotation at the free end and measure the torque. However, I also have no clue how to measure the torque and prescribe the rotation.

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 4 oct. 2011, 01:16 UTC−4
Hi

both problems are linked and can be solved the same way, for the torque I have only the 3.5 examples set up in the model exchange to propose, but this has been discussed often on the Forum, try a search.

Still check the solid.curlUx method for your case, you might get around if you consider the edge, or apply the full formula by hand using tan(alpha) instead of the simplification tan(alpha)=alpha.
I would start to consider to average over the edge the displacement over the distance to the centre of rotation, but where exactly is that located ? depends on your model

--
Good luck
Ivar
Hi both problems are linked and can be solved the same way, for the torque I have only the 3.5 examples set up in the model exchange to propose, but this has been discussed often on the Forum, try a search. Still check the solid.curlUx method for your case, you might get around if you consider the edge, or apply the full formula by hand using tan(alpha) instead of the simplification tan(alpha)=alpha. I would start to consider to average over the edge the displacement over the distance to the centre of rotation, but where exactly is that located ? depends on your model -- 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.