Assembly joints for solids and shells

Feature Requests
1

Predefined joints handled by coupling and equation constraints? Awesome!

2 Likes
Export individual (selected) meshed parts
2

Wow !

Can’t wait to try it.

1 Like
3

I need some more testing, but it looks promising. It uses rigid-body equations and should work for small deformations.

2 Likes
4

Very interesting image. Are the yellow symbols a new type of joint?

5

As mentioned above, they are joints modeled with rigid bodies and equations.

image
image980×728 44.2 KB

image
image674×154 2.72 KB

2 Likes
6

I think their use for shell elements is limited (problems with *Rigid body), and to work at all (CalculiX bugs), some workarounds were necessary, but the tests are promising.

1 Like
7

Did you consider using coupling constraints with orientations instead of rigid body constraints in some cases (e.g. for hinge) ?

image
image794×665 27.4 KB

They could help avoid the known limitations of rigid bodies.

1 Like
8

For solid elements, the rigid joints work. But for shell elements, I tried the kinematic coupling on the shell edges, which crashes CalculiX with the access violation. So I did not investigate it further.

But I did investigate the possibility of using springs instead of couplings-rigid_connections. This is a penalty-based approach that yields approximate results but has no problems with shells/solids. So will probably add it this way.

9

I’ll check it and report to the CalculiX team if needed. Couplings still need a lot of improvements, especially when it comes to their use with 2D elements. But I thought that node-based surface would suffice for kinematic coupling to work with shells.

1 Like
10

I will have to find the example I created to see if I used a node-based surface.

11

It seems to be working with a node-based surface:

Shell edge kinematic coupling.pmx (83.1 KB)

12

I tried your model in CCX 2.22, and it solves, but not in Nlgeom. But it does not solve at all in CCX 2.23. So, for now, I will only support rigid and spring joints. Spring joints work for Nlgeom.

1 Like
13

I moved the conversation about joints to a new topic.

1 Like
14

Yeah, it’s probably the same issue as rigid body + shells + Nlgeom.

That’s strange. A regression in ccx 2.23 ? I have to discuss it with their team then.

15

Did you try a frequency analisys to see if the model is properly constained?. Kinematic needs special care.

1 Like
16

I have to investigate it further anyway before submitting a bug report. This was just a quick test to see if it works at all or crashes.

1 Like