Surface interactions inside another surface

[JohnBobandy]

Hello all,

I have a structure like shown in [1] with a shell extrusion geometry like shown in [2]. And this structure will be axially crushed by a rigid plate (not shown).
As you can see, the structure is multi-celled. But defining the interactions of surface-to-surface (rigid plate to beam) and self-contact (beam with self), only one side of the interior surface can be picked. The same surface is shown at different angles in [3].
If I pick the external sides as expected and only ONE side of the internals, then the simulation is aborted due to the error:

"2 improperly defined surface(s). Please check your surface definitions. Make sure that all surface normals point outward."

And if I check the surface normals, this is what I can see: [4]

I have also tried creating a surface using part->Features->Surfaces and selecting both sides there but the simulation gets a different error in that case:

"Contact pair references surface/node-based surface/analytical rigid surface assembly_beam-1_surf-1 but this surface/node-based surface/analytical rigid surface cannot be used with *contact pair. Check previous warning messages for this surface to find the cause.

Contact pair references surface/node-based surface/analytical rigid surface but this surface/node-based surface/analytical rigid surface has not been defined or this surface is defined on the collapsed faces"

Does anyone have any ideas for solving this? Look forward to reading your responses!

[1]:

[2]:

[3]:

[4]:

 

[Mustaine3]

Why don't you use General Contact? Especially when using A/Explicit.

 

[JohnBobandy]

I have already tried general contact but it produced unfavourable results. I think the interactions are not processed correctly if I use that option:

I'm expecting an axial crush but instead I get something like this:

 

[Mustaine3]

Why do you think that this caused by General Contact?

 

[JohnBobandy]

Well I thought that because the structure should not move like that so it seems like a contact issue. I've varied contact stiffness in this mode a lot and couldn't really get much better. When I did get something better (where a crush happened and no weird movement) the crush stops halfway.

So in that general contact mode, either I get the weirdness above or I get something that barely crushes and then rebounds.

Do you have any suggestions?

 

[Mustaine3]

This has nothing to do with the self contact. Run it without it and you will will get similiar results until the point where the part comes in contact with itself.

I don't know what kind of results you expect. When you think that it collapses without a major lateral deflection, then I have my doubt if the structure isn't too slender for that.

You can try to use imperfections. Do a buckling analysis first and reuse the modes to seed imperfections into your structure. This usually produces much more realistic results in a postbuckling analysis. Look for *Imperfection in the Keyword Reference Manual and also read the chapters that are linked to that. There is also an example in the Example Problems Manual.

 

[JohnBobandy]

The reason I thought this is because a very similar square structure without the "cells" added crushes with the results I expect:

That's why I reasoned it would be the contact that's the issue as that is the only difference between the two shapes (exactly the same length and roughly the same area)

 

[Mustaine3]

As mentioned, work with imperfection.