3D frame with truss beams

seirov

Hello everyone,

I discovered Mecway some weeks ago, and I have found it wonderful. I had also used Lisa (too simple) and Salome (too complex), and it seems that Mecway is just the right size

However, I have a problem that I cannot simulate: I have a frame in 3D built from beams (line2), and I would like to simulate some of the beams as supported (that is, the can rotate at its ends). I know that I can check "truss" from "element properties", but this does not work if the beam is made of several elements (because intermediate nodes don't rotate). Please, see the attachment to make it clear.

I think this might be easily done somehow, as the solution of an element hinged-fixed is known and can be inserted in the rigidity matrix. So, is there a way to simulate this? What have I missed?

Thanks!

prop_design

is the attached file what you are trying to do? config 2 has the rotations at the nodes you wanted suppressed.

edit; updated the attachment

prop_design

here is a slightly different version with a table top attached. just for fun. interesting how it changes the beam deflection though.

edit; updated the attachment

i had to laugh at the results. it's the size of a twin bed but it's saying it can hold 5 m1 abrams tanks.

disla

Hi and welcome,

To allow the rotation on a beam you could also proceed as this:

Activate the thickness of the beam to see its shape.
Select its surface on the extreme like in the picture.
Apply to that surface a flexible joint on beam Boundary condition.
If you only want to allow rotations on a specific plane check on the manual the axis convention on beams.

The attached file shows some different configurations of a portal frame.

If you want more detail on the trusses you can convert it to a regular beam and allow rotation on its extremes the same way as described above. Notice that non symmetrical sections do not model bending twisting coupling.

seirov

Thank you for your help! In fact, the case I meant was what disla proposed. Theoretically, the solution (in 2D, for easy) for an element with both ends fixed is:

and one end hinged and other fixed:


So, the flexible joint was exactly what I was looking for, and I admit that I have never found it on my own (it is a bit tricky how to set it up).

So, I made a simple simulation, to check it with the analytical solution, and it worked like a charm! There were two frames, with and without flexible joints, at the front, vertical beams are not deformed because there is no moment transmitted by the upper beam, due to the flexible joint:



Thanks !

disla

There is something weird on your rotation symbol ¿isn't it?. Do you have the beam correctly oriented ? Your picture shows constrain in the W direction which is the right one but it looks pointing upwards.
If you are getting the right analytical values maybe it is just a visual effect.

Victor

Also interested in what this is in case it's a bug. The deflected view doesn't seem to be consistent with the symbol which is supposed to be oriented like a pin in a pin joint.

prop_design

i think the axis of rotation depends on the element orientation. the attached model is using v as the rotation axis. the pins are aligned the way victor says they should be, in the above post. this model has pretensioned cables and the beams they connect to can rotate about v. i added frictionless supports to the cables/truss elements. this was to prevent free body motion.

the pic, below, shows the pin orientation:



the pic, below, shows what the model looks like:

prop_design

just for the heck of it. i rotated the legs and added more pins. the pin orientation does follow the element orientation. you can enable the element axis in mecway, to see them.

below shows more pins added:



below shows different leg orientation:

Victor

Oh, from looking at @prop_design's model, I think I see what's going on. In the deformed view of the solution, the pin symbols sometimes rotate to funny directions. This is a graphical bug where the rotation of the elements causes their orientations to be recalculated for display of those pins, and that sometimes leads to quite large rotations about the U axis. Thanks @disla, @prop_design, and @seirov for revealing this.

seirov

here there is the "undeformed" view of my frame, where the pin orientation seems to be correct. As you said, it seems a bug in the deformed view.

cwharpe

Also curious how the graphics show the deformed orientation of the joint end faces. One expects a relative angular displacement between faces, but the initial orientation is preserved, thereby distorting the end element of one of the joint members.

However, the numerical results seem to be equivalent regardless of which face is chosen for the pin.

Selecting both faces for pins seemed to default to the single face case.

Victor

Don't read too much into the orientation of the end faces. I think they follow the rotation of the node without regard to the pin so:

When the column is pinned, the beam and node rotate together.
When the beam is pinned, the node remains unrotated with the column.
When both are pinned, the node's rotation isn't part of the system and seems to remain zero in this case but it might be at risk of rigid body rotation.