3D deformable beam self penetrate

[xsong]

thread799-360979

Hello everyone,
I'm doing a simulation: a 3D beam is placed in a rigid box, and compress this box. But there is always penetration between the box and the beam, and penetration also occurs between the beams. I'm using general contacts and tabular for contact properties(normal behavior). These penetrations stop until the centerline is reached.
I don't understand, why penetration occurs, I tried contact pairs, and "Contact controls assignment, contact thickness reduction=self", also doesn't work.
If you can help me with some advice, I'd really appreciate it.
Thank you
XSONG

 

[xsong]

This is the inp file.
Thank you.

 

[FEA way]

Have you considered modeling these wires with solid elements, like in the "Crimp forming with general contact" example problem in the documentation ? In this case the number of wires is not that large and such less approximate approach would make sense. You could also tie surface elements to beams to represent their contact geometry directly.

 

[xsong]

Thank you for your reply, FEA way.
I have try three beams comprsiion and three wires with solid elements, it works.
Sorry, I don't understand "tie surface elements to beams to represent their contact geometry directly".
Would you like to explain it for me?
Thank you very much.

 

[FEA way]

In Abaqus you can model the outer surface of the beam (take just this surface from the solid model), assign surface elements to it, connect it with tie constraints to beam elements representing the wires and run the analysis. Then this surface will be used for contact interactions instead of the default surface computed for beam elements. However, that's quite unusual and complex approach so I would try with solid models of all wires first.

 

[xsong]

Ok, thank you very much.
I will try solid elements.
But, DO you know why the three beam elements works well? The odb figure attached on the reply.

Thank you.

 

[FEA way]

With more wires modeled using beam elements, the spacing might be tighter and it's harder for the solver to properly resolve the contact interactions. You may have to play with various contact algorithms and settings (such as scaling the contact stiffness) to get this to work.