Importing cosmosworks meshes of into nastran

[pandaleone]

HI all

I am trying to import a assembly meshed with Cosmosworks 2005 inside Femap for setting up the analysis with Nastran. I don't want to import the geometry into Femap because it would take a longer time to prepare a good mesh.

In this assembly there are some surfaces on contact (pin/hole, mates, ...) and the resulting mesh contains gaps badly described; for example, for two coincident nodes where the format should be:

CGAP cgap_id pgap_id nodeA nodeB orientation_node

the node orientation is defined as a casual one on the mesh (all CGAPs get the same node for oientation), and therefore does not define the right gap direction. The property PGAP is empty, no values are defined, but this could be left to the user. Moreover, when nodes are coincident, Nastran requires a 6th field in CGAP definition where the CID is defined which Cosmos does not write.

Is there something I'm missing or Cosmosworks translator to Nastran is not reliable for assemblies?

Thank you for your support

 

[GBor]

I've imported COSMOS contact models into FEMAP and run it through NENastran with great success. The problem is not generally in the contact, but rather in the element definition. If you are dealing with shell elements, for some reason they import as plates, but it is easy to define a property for shell elements and redefine all your elements as that property.

I've even done assemblies, so I know it can work, but I'm a little lost on the gap orientation definition and the affect it may have. I would suggest deleting the gap elements and redefining surface contact. It's generally better and would eliminate your gap orientation issue.

Garland

Garland E. Borowski, PE

 

[pandaleone]

I thank you but we should be able to export with Cosmos itself. I just would like to know if I didn't setup Cosmos export or contact options in the proper way or it's bug in the software. I mean: if an exporter to Nastran is provided, I expect that produces the right gap elements.
Has anyone of you had success exporting Nastran or maybe Ansys or Cosmos into FEMAP (which can anyway produce a .nas file for Nastran analysis)?