I am tying to solve...
 
Notifications
Clear all

I am tying to solve a non-linear problem using Ls-Dyna IMPLICIT solver. However, I am unable to achieve convergence.

21 Posts
5 Users
2 Likes
6,354 Views
Posts: 7
Topic starter
(@mohammad_basri)
Unpaid Intern
Joined: 4 years ago

Error - "Non-linear solver failed to find equilibrium"

convergence issue while attempting to compress a specimen with large deformations using IMPLICIT Solver of Ls-Dyna

20 Replies




Negative Volume
Posts: 639
Admin
(@negativevolume)
CEO
Joined: 5 years ago

Hi @mohammad_basri,

To start, Implicit is much more sensitive than explicit so the debugging process can be a long, drawn out process. If you are doing a simple simulation that is in the range of seconds, consider an explicit simulation. If it's longer than that, then yeah, you will have to use Implicit.

Is the issue happening immediately or a little while into the simulation?

 

The link below has a lot of great tips for both debugging errors and improving implicit simulation performance. I would start with the control settings on slide 14 and go from there. 

https://www.dynamore.de/de/download/presentation/dokumente/download-informationstag-Implizit-2016/05-2016-02-dynamore-infotag-implizit-tipstricks.pdf

Let me know if you have any specific questions regarding this. 

Reply
5 Replies
(@mohammad_basri)
Joined: 4 years ago

Unpaid Intern
Posts: 7

@negativevolume

The reason for using Implicit code is to simulate a quasi-static compression problem. 

This error comes after a while and not in the beginning. For a few iterations, it achieves equilibrium and it is after some iterations, it is unable to achieve equilibrium. After which it shows Error termination. 

And, Yes, I have already studied these guidelines for Implicit and used the suggested control settings. 
However, the result is more or less the same. I still get Error termination. 

I am trying to compress a specimen and the process is similar to Crashbox crushing. During the analysis, it shows the following error:

"divergence (increasing translational residual norm) detected"

Reply
Negative Volume
Admin
(@negativevolume)
Joined: 5 years ago

CEO
Posts: 639

@mohammad_basri

If you are getting an error in the middle of your simulation, it could be an issue with the material model or the contacts. Are there any contacts in this simulation? If so, that could definitely be the cause. This link may be helpful

https://www.dynasupport.com/howtos/implicit/implicit-contact-treatment

From my experience, implicit simulations require more than just baseline contact settings so you are probably going to need to activate the options ABC in the contact in LS-Prepost (below) and set IGAP to 1. IGAP has a huge effect on implicit contact stability. 

 

 igap contact
Reply
(@mohammad_basri)
Joined: 4 years ago

Unpaid Intern
Posts: 7

@negativevolume

I have already given and tried the IGAP =1 in Contact Card. 
I am currently using Mortar Contacts. 

I will try changing the contacts once. 

also, do you have some idea about the control card CONTROL_IMPLICIT_SOLUTION? Since the solving process os governed by this itself, I believe there is something missing in that itself.

Reply
(@ssdng)
Joined: 2 years ago

Barista
Posts: 3

HI @negativevolume 

Could you share the pdf file to which the link you have shared? I don't have access to the site. And I really have an issue with non convergence of my implicit analysis

https://www.dynamore.de/de/download/presentation/dokumente/download-informationstag-Implizit-2016/05-2016-02-dynamore-infotag-implizit-tipstricks.pdf

Reply
Negative Volume
Admin
(@negativevolume)
Joined: 5 years ago

CEO
Posts: 639

@ssdng Hmm, it seems the webpage that I originally linked is either down or no longer exists. Let me see if I can track it down for you. 

Reply




Posts: 7
Topic starter
(@mohammad_basri)
Unpaid Intern
Joined: 4 years ago

Can you have a look at my Message file and suggest changes in the CONTROL_IMPLICIT_SOLUTION?

Reply
4 Replies
Negative Volume
Admin
(@negativevolume)
Joined: 5 years ago

CEO
Posts: 639

@mohammad_basri

Did you send the message file to my email? I’m sorry, this slipped by me.

Reply
(@mohammad_basri)
Joined: 4 years ago

Unpaid Intern
Posts: 7

@negativevolume

I had actually uploaded the file. But somehow it didn't. Can you share your email id?

Reply
Negative Volume
Admin
(@negativevolume)
Joined: 5 years ago

CEO
Posts: 639

@mohammad_basri

Just change the extension of the file from .msg to .txt and upload it. 

Reply
(@mohammad_basri)
Joined: 4 years ago

Unpaid Intern
Posts: 7

@negativevolume

I am attaching the message file. 

Reply




Posts: 5
(@delcon)
Unpaid Intern
Joined: 3 years ago

@negativevolume

@mohammad_basri

Hi!

Tell me pls, was the problem solved? Suppose I have something like it.

Reply
1 Reply
(@mohammad_basri)
Joined: 4 years ago

Unpaid Intern
Posts: 7

@delcon No, the problem is still there, what is your issue?

Reply




Posts: 5
(@delcon)
Unpaid Intern
Joined: 3 years ago

@mohammad_basri 

I am solving some simple along/across fiber tensile problems for composite. 
Almost all of them (34 of 35) were solved ok, except 1 (so the problem i suppose in material card or contact - 1 tied to node to surf contact)

This material (mat_054) fails before stress riches XT (xt=518Mpa, fail on 350-400Mpa).

According to logs timestep falls below DTMIN (control_implicit auto) . But no idea why.

I suppose the problem is with convergence, but no idea how to fix this.

And I also got this error "Non-linear solver failed to find equilibrium"

Reply




Page 1 / 2




Share: