Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.
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.
Torsion Simulation in 2D axisymmetric
Posted 16 set 2012, 19:56 GMT-4 Materials, Modeling Tools & Definitions, Parameters, Variables, & Functions, Structural Mechanics Version 4.3 11 Replies
Please login with a confirmed email address before reporting spam
I am going to simulate a rod using the structural mechanics module in COMSOL v4.3. The rod is fixed at one end and a torque T is applied in the other end. So the strain is pure torsional.
Can I use 2D axisymmetric to simulate this rod?
If yes, how could I apply that boundary torque? According to material mechanics, I should apply an out-of-plane stress on the boundary, but I can only find Tr (radial) and Tz (axial) when I add a "boundary load".
Thanks
Dan
Please login with a confirmed email address before reporting spam
Good quastion, I assume so that 2D-Axi allows for a pure axial twist, but you would have to tweak a little the equations.
In 3D it's clearly taken into account. The only thing is that to apply a moment you need to use weak conditions, or use a "Rigid Connector" BC, but this changes also the boundary to a"rigid" part, as the "soft connector" is not yet implemented, but I have heard it's being prepared ;)
There are some eamples in the model exchange, even if for older COSMSOL versions, but you should be able to translate these
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Please login with a confirmed email address before reporting spam
thanks for the precision, so I take by the "tweaking" in 2D-axi, it would be reather rewriting the approach ;)
so far simpler to go to 3D, as there it works nicely
--
Have fun COMSOLing
Ivar
Please login with a confirmed email address before reporting spam
No, you cannot use the 2D axisymmetric environment, if for no other reason because it presumes that all displacements are in the plane, which yours clearly aren't.
If this method doesn't work, can I apply other ways such as PDEs in COMSOL to make use of the structural symmetric?
for an axisymmetric geometry, the stress-strain relation can be written as:
[Trr;Tzz;Trz;Tphi] = [E] * [Srr;Szz;Srz;Sphi], the torsional strain Tphi is inside this cylinder cooridnate. Can I set up a 2-D axisymmetric PDE weak form model to solve this problem?
In COMSOL 2D axisymmetric, displacement should be in-plane, where are the Tphi and Sphi components?
Thanks a lot
Dan
Please login with a confirmed email address before reporting spam
turn on the equation view (options preferences) to see some of the inside equations, but with the tensor notation in 2D-axi you need to take some care ...
Why not use 3D ?
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Hi
turn on the equation view (options preferences) to see some of the inside equations, but with the tensor notation in 2D-axi you need to take some care ...
Why not use 3D ?
--
Good luck
Ivar
3D model is too much slow for my case. I think there might be a way to make use of the geometry symmetry. I'll check the inside equations right now. Thanks
Please login with a confirmed email address before reporting spam
The following relation transfer the 3-D model to 2-D:
dV = 2*pi*r*dS, where dS = dr*dz
After this simulation, phi is no longer in the integration.
Do you think it's more efficient to use cylinder cooridnate than x-y-z for torsion simulation?
Please login with a confirmed email address before reporting spam
its easy to add a cylindrical coordinate in 3D, and then you can use the sys2.r instead of sqrt(x^2+y^2)
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Apart from doing a full 3D simulation, there are two more efficient options:
1. Create a 3D model with only a sector (say 30 degrees) of the the structure. Use periodic boundary boundary conditions (the Periodic Condition feature).
2. If you feel confident with the theory, augment the built-in 2D axisymmetric physics with an extra degree of freedom for the twist. Then set up variables for the new stress and strain components and add them to the weak (virtual work) expression.
Regards,
Henrik
Please login with a confirmed email address before reporting spam
Also keep in mind that for transient/dynamic problems the deformation due to twist/torsional loading is frequently not axisymmetric.
Nagi Elabbasi
Veryst Engineering
Please login with a confirmed email address before reporting spam
Starting at version 6.0, torsion of axisymmetric objects can be modeled in the 2D axisymmetric interface by selecting the "Include Circumferential Displacement" option ( This follows the idea explained by Henrik above, now available "out of the box"). Read this blog post for more information on this topic.
Best regards,
Jeff
-------------------Jeff Hiller