Rigid Link/connection - Karamba v. 1.3.0


(Manta Leonardo) #1

Hi Guys, I need your help for a simple question that I’m not be able to solve: I need to connect a reticular element to the section of a steel beam, and to do that, I want to create rigid link, so I created a line element from the end/start point of the beam (as in picture attached), but I don’t know how to make them follow the section beam deformation. Moreover if I don’t connect a cross section to this vertical element, they don’t move and stay in linear position.
I also read about the possibility to use springs, but I really don’t understand how can help me in this case.
Hope someone can help me.



Dear Leonardo,
in Karamba3D elements are connected to each other only when they have a common starting- or end-node. This does not seem to be the case in your definition between the diagonals and the I-beams which support the slab.

(Manta Leonardo) #3

I know because I’m not able to realize a rigid vertical link for the diagonals. So, how can I connect the diagonals to the top and bottom of the I-beam? Am I doing it right with the rigid connection element or are there some other tips I can use?

Thank you!



You would have to e.g. add two small beam-pieces which start from the endpoint of the large I-profile and end at its flanges. For this you could use springs. Generate springs by adding a spring cross section to the beam. Use high stiffness values for the translational and rotational dofs of the spring and test different values. If you use a stiffness which is too large, the stiffness matrix may become ill-conditioned, which leads to inaccurate results.

(Manta Leonardo) #5

Thank you so Much Clemens. I’ve already done the small beams at the start/end of the I principal Beam, but I didn’t know I’ve to use springs. I’ve another question: also if I used the spring cross section with high stiffness Ct and Cr, the point at the end of the flanges seem not connected with the diagonals as I want. I’ve noitced that the effective connection between I-profile Beam, the small beams and the diagonals, performs only if I use a stronger section for the small beam (es. IPE 600), but I know this is incorrect.
What Am I wrong? I attached a photo with the springs use and the example with IPE 600 for small beams.

Thank you so much for your help, I really apprecciate it.


Dear Manta Leonardo,
could you send me a simplified version of your definition with only one or two connections that do not work. This would make it easier to spot the problem.

(Manta Leonardo) #7

Sure, The model I provided is parametric so I deleted beam and diagonals to the minimum number in order to analyze what is going wrong. Let me know if it’s still too element to identify the problem, so I will create a single element in separate file.
I group the small beam section in the file so you can easily identify them.
Thank you so much! I hope to solve finally this problem.


Karamba Test.gh (51.2 KB)

(Manta Leonardo) #8

Hi clemens, have you seen the attached file? I also try different alternatives to connect the element but I failed.

I need a solution to go on my project. If you can help me, please.



Hi Leonardo,
I have tried a similar example (see KarambaTest_cp.gh (45.5 KB)).
My first finding was, that a spring is not the right connection element in this case, since it connects two nodes without taking account of the kinematic relation (see manual).
When using a beam element it works better, but there is still a gap which opens with increasing scaling factor for the displacements. The reason is, that the nodal displacements are calculated assuming small displacements whereas the transformation of the beam mesh is done on the basis of large rotations.

(Manta Leonardo) #10

Sorry for the late response. Anyway thank you for your explanatin Clemens, I’ve assumed it like you said.