Discussion Groups

Simulation Mechanical and Multiphysics

Reply
New Member
roypersa
Posts: 2
Registered: ‎11-01-2012

Convergence issue with elastic-plastic material... help needed!

339 Views, 3 Replies
11-01-2012 02:41 PM

Dear all,

 

I've been struggling with a von Mises nonlinear 2D axisymmetric problem for a while.

Geometry resembles a simple closed cylindrical shape. Apart from symmetry at two edges (one quarter model), the unique load is pressure. I've played around with almost every analysis parameter of the Advanced tab including increasing the displacement tolerance to 0.001. Since it's a quasi-static problem, I'm using static with nonlinear material instead of MES. Newton-Raphson 1 without line search has so far given thecloser to the end result but yet unconverged.

 

Are there any guidelines you can share about convergence enhancement?. By the way, model seems to be correctly set-up since using a linear material model it converges quite easy. The pressure load utilized responds to that physically utilized in the test.

 

Kind regards,

 

Roger

 

 

Please use plain text.
Employee
zhuangs
Posts: 261
Registered: ‎10-04-2010

Re: Convergence issue with elastic-plastic material... help needed!

12-12-2012 10:29 AM in reply to: roypersa

You might share the model for us to have a look.

 

Shoubing

Please use plain text.
Active Member
pbuscemi46
Posts: 10
Registered: ‎05-01-2010

Re: Convergence issue with elastic-plastic material... help needed!

01-04-2013 09:10 AM in reply to: roypersa

 itself  Without seeing the model and its parameters its difficult to offer suggestions. but...

 

1) Try some smaller mesh sizes,  2 )  check to see that strains are small so that non-linear static really applies, 3) don't fret using MES. it's all that I use   Decreasing the tolerance may be a factor and suggests- if it helped at all , that MES is the better approach

Please use plain text.
Employee
zhuangs
Posts: 261
Registered: ‎10-04-2010

Re: Convergence issue with elastic-plastic material... help needed!

02-05-2013 12:39 PM in reply to: roypersa

How many parts are there in the model? Is there surface contact used?  If there is surface contact, w/o line search might has trouble for convergence.

 

Is "Vos Mises with Isotropic/Kinematic Hardening" or "Vos Mises Curve with Isotropic/Kinematic Hardening" used? For both of them, finer mesh with good mesh quality is needed for better convergence and solution.

 

-Shoubing

Please use plain text.