-
-
November 22, 2023 at 9:48 amVincent RademacherSubscriber
Hello,Â
I am currently trying to perform a Full Restart on a model. So far everything works but there is one problem: I remove some parts from the original model and add new parts in the restarted simulation. The contacts assigned to the removed parts are killed during the original simulation via birth/death. When I try to run the restart with these killed contact cards removed or new contact cards introduced the calculation is not started and no error message is printed. The modelcheck (.mchekout file) stops after the line "initialization completed" not reaching the "normal termination" line.Â
Currently a workaround is possible: I use the ID´s of the contact cards from the original model (contcats already killed during original simulation) and just assign new master and slave IDs from the new included parts to them. This only works if all contact ID´s present in the original simulation are also prensent in the restarted simulation. That seems wrong to me as those are entirely new contacts with new parts. I would think that it should be possible to remove contatc ID´s from contacts no longer used and introduce new contatcs with new ID´s? At least that is what LS-Dyna 12.0 documentation suggests:Â
"Pre-existing contacts, in general, carry forward seamlessly using data from theÂ
d3dump (or d3full if MPP) database. It is important that the contact ID(s) in theÂ
full restart input deck match the contact ID(s) in the original input deck if theÂ
intent is for the contacts to be initialized using data from the d3dump/d3full
database."
My intent is to remove contacts and add new contacts just initializing those that are present in both simulations. I am happy for any suggestions as the workaround will fail once I am working with models that will have an unequal number of removed and added contacts or situations where no contacts are added at all but parts are removed.Â
Kind regards, Vince -
November 22, 2023 at 12:53 pmVincent RademacherSubscriber
Came up with a solution myself just now. The problem seems to be that only those contacts that should exist in both simulations (original and restart) are to be defined with the "_ID" option in the contact card. Prior I had used "*CONTACT_AUTOMATIC_SURFACE_TO_SURFACE_ID" for all contacts in the original simulation. Not mentioning all the IDs in the restarted simulation resulted in the job not being submitted properly. Now I use "*CONTACT_AUTOMATIC_SURFACE_TO_SURFACE" for all contacts that are not carried over to the restart simulation aswell as those that are newly created in the restart simulation. Everything works as expected.Â
Maybe the dodumentation should be somewhat updated if this problem is not limited to my specific model/machine/version. The way I understand the excerpt above it is not clear that I have to reuse all contact IDs that are defined in the original model.Â
-
- The topic ‘Contacts in Full Restart’ is closed to new replies.
- LS-DYNA Installation Issues with Student Workbench 2024 R2
- LS-Dyna CESE SMP d vs MPP d solver
- Cross-coupled stiffness elements in LS-DYNA
- CESE solver – Ignition mechanism
- Mathematical model generation stuck at 10%
- About combine different unconnected body into one part
- Tiebreak using Segment set for contact b/w 20 noded Hexahedral elements
- shape memory alloy material in LS-DYNA
- CESE combustion model
- Initial Stress Shell Application and HistVarCosine in LS-DYNA
-
1301
-
591
-
544
-
524
-
366
© 2025 Copyright ANSYS, Inc. All rights reserved.