No Convergence on Sol106 MSC.Nastran

[tobiknobi]

Hallo,

My problem is, that my Modell is not converging.

I generated a 3D Modell of a double lap shear rivet connection. Rivet an Sheet are modelled with Hex8 Elements. Between the Rivet and the Sheet is a small gap. The connecting nodes there are connected with CGap Elements. (Because MSC.Nastran doesn't provide 3D Contacts)

When I run this nonlinear, I always get convergence problems. Because of excessive pivot ratios at the Gap Element nodes on the Rivet. The rivet is free in sheet directions, but constraint, rotational constraint normal to this direction.

Some of the Parameters, I used are:

NLPARM,1,10,,AUTO,5,25,UPW,NO
.01,.01,.01
PGAP,3,.01,,5.+7,.005,5.+6

Thanks
Tobias

 

[JacobL]

you can manipulate the nlparam options for convergence, but in general the more complicated the model, the higher chances you have of not making it work.

have you tried a finer mesh?

Jake.

 

[tobiknobi]

Hey JacobL,

no, I didnt try a finer mesh. but the mesh is very fine already. And I played alot with the nlparam values.
What I found out recently is, that a 3D contact cant be done with MSC.Nastran. (it is hard to model a 3D contact with 1D elements)
So I changed to Ansys.

Thx 4 the post

 

[JacobL]

actually, for nonlinear solutions, msc marc is the answer..

nastran is not really good at non linear question.