how to define boundary condition along surface normal direction

[jisb007]

Hi,

I have a curve-shaped object that I am modeling to have the boundary nodes fixed in the surface normal direction, while allowing tangential movement. How can I specify such a boundary condition then? I looked up "*TRANSFORM" keyword, but it seems that the local coordinate system is very simple, only Cartesian, cylindrical, or spherical. I don't want to specify a local Cartesian coordinate for each of my boundary node. Maybe this is not the way it should be handled.

Someone knows how to do this? It's probably very simple, but I just haven't got the answer. Thanks very much!

 

[portliner]

There is an option "allow separation after contact" on the contact property > mechanical > normal behaviour. Maybe this can be used by deselect it.

 

[jisb007]

Thanks for the tip, but that's not the case for me. I don't really have contact definition involved -- all I need is to specify that the nodes on the outer boundary can only move tangentially, but not in the normal direction.

 

[rstupplebeen]

Typically you will get better answers if you describe what you are trying to model with pictures or CAE files then pose your question. With more information there may be a simpler or more accurate solution.

For instance what shape is the surface? If this is a hip joint then a simple radial constraint may be appropriate...

I hope this helps.

Rob Stupplebeen

https://sites.google.com/site/robertkstupplebeen/

 

[jisb007]

Hi Rob,

Thanks for your reply -- unfortunately, I do not have a CAE file to show what it looks like as I generate Abaqus input files directly without using CAE.

The question is more of a generic in nature -- I know when implementing the finite element code for analysis, it's actually quite easy to specify normal/tangential boundary conditions for boundary nodes (i.e., multiplying a rotational matrix when assembling into the stiffness matrix). However, just don't know how to do this in Abaqus.

The specific problem I am trying to solve is to model the human brain -- I wanted the brain boundary nodes to be able to move in tangential direction, but fixed in the normal direction. The shape is more complex than a simple sphere, as the nodal normal direction differs at different location.

Can you please elaborate a little bit more on "radial constraint"?

Thanks!