Thin shell assemblies, bonding

[CalculatorDave]

Howdy all,

I have an Assembly with parts have fail to bond during FEA:

I have Cosmos Professional 2006. I have built an Assembly from Parts. Each Part is made from zero-thickness surfaces. I assembled the pieces, mated them, and ran a linear statics analysis. Some of the pieces fell away from the assembly. I have NOT tried to enforce mesh-matching through the use of dozens of split lines. I HAVE tried, without success, to use Contact/Gaps to join the components (maybe this only works for solids?).

There's GOT to be an easier way to do this. Why doesn't COSMOS "understand" that, if two surfaces are coplanar and in contact with one another, they are BONDED?

Any advice?

David

 

[tnkayaker]

From my limited experience using Cosmos with assy's I've not found an easy way around the split lines when doing shell meshes.

 

[cbrn]

Hi,
as a general rule, make sure you don't confuse CAD's constraints with FEM's constraints. CAD constraints will NOT become FEM restraints because they are conceptually different. If you have a very complex assy, unfortunately, YES, you will HAVE to create FEM matings between the parts. These "matings" will have to be "realistic".
Please understand that there are many many situations where two co-planar or co-axial surfaces are NOT bonded, so that's an assumption that a FEM MUST NOT DO BY ITSELF. Also, "coaxial" is a coherent definition from a CAD point of view, but almost meaningless for a FEM (or, at least, not sufficient): is it a pin around which the parts can rotate? does it have friction? does it have axial preload? and so on...
Using "coincident meshes" at the interfaces btw parts will allow to "bond" them together, but in many cases will cause mesher impossibilities (this has already been discussed in a thread some time ago). However, you could take advantage of CW's "connectors".

Hope this helps...

Regards