How can I properly constrain a 3D truss beam in ANSYS using only beam elements?

[AJ Kazakov]

Hello!
I'm trying to model my truss with FEM using only beam elements in ANSYS, but I am experiencing difficulties with the constraints between the members (or coupling of the DOFs).
For instance, I want my truss chords to be continuous and my lattice members to be pinned to them. I couple all of the translational DOFs, but don't know what to do with the rotational ones.
I would appreciate any help with that.

 

[paisiello2]

How do you couple all the translational DOF's?

 

[AJ Kazakov]

I set the nodes of the different elements in a single joint to have the same translations e.g. two diagonals and a vertical meet the chord at the same joint, the chord is rigid (continuous), and the end nodes of the diagonals and the vertical have the same translations as the node of the chord at that joint. But what about the rotations? I don't want to overconstrain the lattice members, but make them pinned to the chord.

 

[nvn]

AJ Kazakov: Perhaps try this. When you merge chord keypoints, deselect lattice member keypoints. I.e., do not merge lattice member keypoints. Then, create coupled dof, for translations only, between the lattice member nodes and the chord node (?).

 

[AJ Kazakov]

AJ Kazakov: Perhaps try this. When you merge chord keypoints, deselect lattice member keypoints. I.e., do not merge lattice member keypoints. Then, create coupled dof, for translations only, between the lattice member nodes and the chord node (?).

This works with trusses in 2D.
But let's consider this truss in 3D:

http://imagizer.imageshack.us/v2/800x600q90/839/pg60.png

What rotational coupling should I have for my lattice members with the chord?

 

[nvn]

But let's consider this truss in 3D. What rotational coupling should I have for my lattice members with the chord?

None. You should have no rotational coupling (about the x, y, and z axes) of the lattice members with each chord.

 

[AJ Kazakov]

None. You should have no rotational coupling (about the x, y, and z axes) of the lattice members with each chord.

But aren't the lattice members going to rotate about their own axes then? The same goes for the middle chord. How to constrain them against that?

 

[nvn]

AJ Kazakov: Good point. I currently am not sure how to do it in Ansys. In other programs, I probably would release the beam element local Ry, Rz dof at end 1, and release Rx, Ry, Rz at end 2, on each lattice member, where, e.g., Rx = element rotational dof about the element local CS x axis. For the middle chord member, you could either torsionally constrain it to ground on one end, or else you could put a light torsional spring (k_x = 1 N*mm/rad) to ground on one end.

Does Ansys have a beam element option to release beam element end dofs, perhaps called beam releases, or beam end releases?

 

[AJ Kazakov]

I have found one element in ANSYS - MPC184 (Multipoint Constraint) which can simulate a joint with a local coordinate system. I'll give it a try.