with RESI_GLOB_RELA too high contact convergence problem

vishwas

Hello,
I am tring a tube forming simulation. The geometry is as attached below. It has a pipe and 3 rollers (front, rear and bot)

The goal is to deform tube in between rollers, then rotate rollers to continuously deform and feed or pull in pipe something like below(done in nx nastran)

link(https://www.youtube.com/watch?v=KbEx2gU4tPw&t=1s)
The bottom roller. moves up to deform the tube held against two other rollers.
Meshing in all hexas and quads,

without friction, I had success till 0.25s but if friction is acrtivtes, none of contact methods(STANDARD or PENALISATION) works.
I also tried with only shell mesh for roller geometry(rear,bot, front) but it gave completely bad results. It resulted in this deformation.

So I had to extrude all of shell mesh along normal with only 1 layer to create hex elements.(thus an all solid mesh)
All study files here: https://www.dropbox.com/scl/fi/w3jva02grg5bvdot34g1l/tube_roll.tar.gz?rlkey=1slvl7yd68v9yu8bu1e2air0w&st=w7chiogb&dl=0
Also the performance of iterations is slow and increasing no. of cores by mpi_nbcpu does not help. It seems that contact friction cannot truly benefit from MPI parallelism. I channged from 4 to 8 to 20 cores but not much speedup could be observed.

It would be much appreciated if someone takes a look.

P.S: I also faced issue with "reload from disc" operation in salome which I posted here:https://discourse.salome-platform.org/t/transfer-mesh-based-on-imported-step-geom-after-using-reload-from-disk-cause-errors/3404/2
link for geom and other files for this prob:https://www.dropbox.com/scl/fi/fy6408nu6outdmhryhmb9/study.tar.gz?rlkey=8u2kew7qgjcri73h6cli5v2u6&st=tz44zbz9&dl=0

Any help is appreciated for above questions.

Thanks and Regards
Anirudh

vishwas

Hello,
I consistently get the following type of error:

[1,2]<stdout>:  POINT A PROJETER :   -8.0000000000000002E-002  -6.1268873859686244E-003  0.11115704824297046  
[1,2]<stdout>:  MAILLE             QU4                         4
[1,2]<stdout>:    NOEUD                     1
[1,2]<stdout>:     (X,Y,Z)  -7.0661384731361798E-002  -2.7976792085603029E-006   1.0565750273932483E-002
[1,2]<stdout>:    NOEUD                     2
[1,2]<stdout>:     (X,Y,Z)  -7.4589953634312686E-002  -2.7292115170730509E-006   1.0239378969976374E-002
[1,2]<stdout>:    NOEUD                     3
[1,2]<stdout>:     (X,Y,Z)  -7.4607753450048636E-002  -1.2220277036085177E-003   1.0206658758570629E-002
[1,2]<stdout>:    NOEUD                     4
[1,2]<stdout>:     (X,Y,Z)  -7.0648071610678917E-002  -1.2366097851533786E-003   1.0485484072487509E-002
[1,2]<stdout>:  KSI   :    3.3260704719549468        12.105876567313597     
[1,2]<stdout>:  ALPHA :    1.0000000000000000     

 [1,3]<stdout>: <Erreur> Échec de l'appariement de contact lors de l'actualisation géométrique en Newton généralisé.
[1,3]<stdout>:  
[1,2]<stdout>: <Action> On essaie de découper le pas de temps.

both contacting surfaces are curved.the dies are curved in two direction. does that cause this error?

Where can I get more info regarding this pairing problem?
Regards
Anirudh