converting the rigid connections to pinned in a space frame

Hello!

I want to run Karamba on a set of geometries (space frames, polyhedra).

Now the lines of these geometries are rigidly connected. But I would like to have pinned connections instead (but no truss system, I want to have bending in the elements), which cannot transfer bending moments.

I know that I can do this with "Beam-Joints". But I can't select all beams and modify the start and end joint accordingly, because then I will have a mechanism (because to much of dof are released in one node).

Is it like this?

If there are 3 beams coming together in 1 node (let's say node A), I have to change the correct start-or-end node (related to node A) of 2 beams? If there are 4 beams coming together in 1 node (node B), I have to change the correct start-or-end node (related to node B) of 3 beams?

Because in this way it seems that there is still a moment transferred in the connections nodes.

Or do I have to do this like described above but for all beams coming together in the node?

Thanks!

Lara

Replies to This Discussion

Permalink Reply by Francesco Ranaudo on May 30, 2013 at 10:01am

Hi lara,

i'm not sure that i've understood very well your question, but if you want to make a pinned connection you can use this definition:
http://www.grasshopper3d.com/group/karamba/forum/topics/zero-length-element-as-spring?commentId=2985220%3AComment%3A644487&xg_source=activity&groupId=2985220%3AGroup%3A251211

Permalink Reply by Lara Alegria Mira on May 30, 2013 at 11:00am

Thank you. I indeed know about the scissor concept. But for this problem it is preferably to use the beam-joints component (which is basically similar).

The problem is that I don't know which beams in a node, where there are more coming together, to release and how..

Permalink Reply by Francesco Ranaudo on May 30, 2013 at 11:24am

ok...tell me if i get the point: you need to make a 3d truss-like structure in which elements can be bent and the nodes can't transfer bending moments but when you set all starts and ends releases you get a mechanism. is it right?

if it is so, do you have a rigid pattern or not? because if you free all rotations in the nodes you need to have a rigid scheme to avoid internal lability.

if your pattern is rigid, probably you need to check boundaries conditions.

try to post some images of what you would like to do, so it will be easier to understand

Permalink Reply by Lara Alegria Mira on May 30, 2013 at 11:59am

You got it right. First question: yes!

I do have a rigid pattern, the beams are in a triangulated formation: they are forming an icosahedron dome.

I have 5 supports, all rigid (all 6 DOF are fixed).

With this setting and releasing all rotations (at end and start points), the analyse component still turns red. (fig1)

When I release only Ry and Rz both at the end and start, then the analyse component works, but then I still have bending in some nodes (fig2-3).

Attachments:

fig1.png, 506 KB
fig3.png, 531 KB
fig2.png, 531 KB

Permalink Reply by Francesco Ranaudo on May 30, 2013 at 12:20pm

ok...probably (but i'm really not sure of what i'm saying, so be kind with me!! :D ) it is because when you set all rotations free, even if you've got a support, this is disconnected (see the manual at page 42). try to release the rotations in all the support-nodes exept one...if i'm right it should work. let me know!

Permalink Reply by Lara Alegria Mira on May 30, 2013 at 1:07pm

Yes, that is what I thought as well. But then I am stuck with another problem: see discussion "Sets of Beams".

RSS