foam structure under compression with self contact

[lifepaper]

Dear all,

My current research involves the simulation of a foam structure (3d voronoi) under compression.

at this stage I have 2 problems,

1.In order to take torsion, bending, buckling etc into account, I intend to choose 'beam' as the section. However, even for a circular cross section, an orientation has to be assigned to each of the beam member. In the case of a foam, each member has a different orientation. Assigning orientations manually to each of the member seems impractical. So how to tackle this problem?

2. Once compression is applied, there will be severe deformation and consequently the self contact of the foam. How to simulate the interactions between these beams?

Thank you.

 

[adfergusson]

Actually, you have another more important problem; your geometry is neither realistic nor periodic. For a start, those vertical beams that define the corners of you model are going to totally dominate the reponse of this structure under a vertical load and real foams do not have such uniform features.

In order to simulate a bulk material from a small volume of it one would typically look to create a statistically representative volume or perhaps a unit cell (if the material is regular, e.g a honeycomb core or hexagonally packed spheres) that has periodic geometry. This video clip gives you an idea of why one might want to use periodic geometry and boundary conditions (Silvestre Pinho may be best known as a composites guy, but he does sandwich panels too and the principals are very broadly applicable anyway):Link

Developing the maths and then implementing it in code to create a periodic voronoi cell will be a challenge unto itself; look here for some ideas: [URL unfurl="true"]http://www.cgal.org/Events/PeriodicSpacesWorkshop/talk_manuel_caroli.pdf[/url]

 

[IceBreakerSours]

If you really need to model those discrete beams, I think connectors should be very useful in this situation. You can also model interaction between connectors (i.e., mechanical contact) as well at (relatively) low computational expense.

However, I would prefer adfergusson's suggested route.

*********************************************************
Are you new to this forum? If so, please read these FAQs:

http://www.eng-tips.com/faqs.cfm?fid=376

http://www.eng-tips.com/faqs.cfm?fid=1083

 

[Mustaine3]

I would throw in a tet mesh, and run it with A/Explicit. Activate general contact. If the deformation speed is low, it might be useful to use some mass scaling to save computation time.


@IceBreakerSours:
I don't think that connectors can have contact.

 

[IceBreakerSours]

You are right, Mustaine3. It turns out, with connector elements, interaction is limited to kinematic connections (e.g., in a slot) - which is what I had used them for.

*********************************************************
Are you new to this forum? If so, please read these FAQs:

http://www.eng-tips.com/faqs.cfm?fid=376

http://www.eng-tips.com/faqs.cfm?fid=1083