Hi Demian!

On 04/19/2013 08:45 PM, Demian Wassermann wrote:

Hello SfePy-ers!

I'm starting to work with the tool. I found it great, very nice to use.

Thanks!

I have a problem that I can not wrap my head around:

I have a cylinder close to a sphere, both with different elasticity properties (the cylinder is rigid and the sphere is more elastic).

I'm trying to rotate the cylinder about the centre of its bottom cap and thus hit the sphere deforming it.

It sounds like a contact problem - the code cannot solve contacts yet - are you trying some workaround?

To do this I put the sphere and the cylinder close, I put them in a bounding box representing the media in wich both are inscribed (air, water, etc), I calculate a mesh for this domain with tetgen having a different material for each object (cylinder, baloon, media)

The medium is also meshed?

and then use a function in the ebcs conditions to move the rod. Similar to the large deformations hyperelastic example (rotate_yz), where the rotation angle increases with time.

My problem is that if my cylinder rotates more than ~.1 degrees, I get a "warp violation" error.

If the medium is also meshed and deforms, then my guess is that it's much softer than the two structures and deforms too much in a single step.

How could I solve this? By making the elements of my mesh smaller? Taking smaller steps, increasing the order of the balance integral holding the elasticity properties?

Yes, smaller time steps could help.

Can you send me the example (off-list if it's big) so that I can have a look?

r.