Contact Simulation with a Pipe and a donut shaped object

Saleh

Hi everyone,
I am trying to do a simulation where a donut shaped object hit a pipe shaped object. I want to define a contact with them. I have been trying to do that by making both of them solid 3D or one solid one shell type.
but the problem has been exit code 137,139 and 4.
I have been using salome meca with virtual box. Now I am using it by Salome for Windows. Can Anyone help?

Bonjour à tous,
Je tente de réaliser une simulation où un objet en forme de donut frappe un objet en forme de tuyau. Je souhaite définir un contact entre eux. J’ai essayé de le faire en rendant les deux objets solides en 3D ou en utilisant un objet solide et l’autre de type coque.
Cependant, le problème que je rencontre est lié aux codes d’erreur 137, 139 et 4.
J’utilisais Salome-Meca avec VirtualBox, mais maintenant je l’utilise sur Salome pour Windows.
Quelqu’un pourrait-il m’aider ?

Thanks
Saleh

mf

Saleh
Hello,

ideally, you should upload an example somebody in this forum can work on. It doesn't matter much if it works or not. But your question above is just too generic,

Mario.

Saleh

mf
Hi Mario,
here is the mesh files and the comm file.
can anyone help me

Saleh

[https://drive.google.com/drive/folders/1JXfEHHhk8K_H2UWC2Y7WWlDypn5jkykn?usp=sharing]

I could not upload the mesh files and zip files. I am providing a drive link.
I am using a virtual box, salome meca 2019.0.1

jeanpierreaubry

Saleh
hello

the way you are doing things is very strange
you look like trying to glue together all nodes of the pipe to all elements of the torus
this way you are creating a solid block with all the elements
no wonder the code gets lost somewhere

if you want to simaulate the two objects to come in contact you have to use DEFI_CONTACT with a STAT_NON_LINE analysis

besides that the meshes seem to me far too refined
i would make a first try with element size between five and ten times larger to test the problem

Saleh

jeanpierreaubry
hi,
I have already tried with different types of mesh, that too with coarse mesh and finer mesh. I have been stuck with this for more than 2 months. I did also used the DEFI_CONTACT also.
Actually my intention is to see the deformation of pipe due to the load given by torus object. Torus must be in contact as I do not want impact load.
Also I do read your book Jean. Thank you for your guidance.

jeanpierreaubry

Saleh
is this the type of behavior you expect
if so
i will develop

Saleh

jeanpierreaubry
hi,
Yes I do want a similar type of behaviour

Saleh

jeanpierreaubry
Also it can a 3D and 3D object contact too

jeanpierreaubry

Saleh

Also it can a 3D and 3D object contact too

i am not sure to understand what you mean
is this a statement or a question?
from the included picture it seems you solved the problem

Saleh

jeanpierreaubry
hi,
3d and 3d was a statement, I meant instead of a shell pipe we can use 3d pipe.
Also I cannot solve it. This is just a picture. I still need help

jeanpierreaubry

Saleh

here is a link to an archive with the problem's files
there is a readme file
valid 48 hours
https://transfert.free.fr/yvNif8Z

Saleh

Hi kind users of Code_Aster and Salome-Meca,
I am again trying to simulate a pipe indented by a torus-shaped object. This time I decided to model the pipe with shell elements.

Shell model choice (COQUE_3D + non-linear material)
I am currently using COQUE_3D for the pipe with a QUAD9 mesh. The pipe material is non-linear, and I have a stress–strain curve from an Excel file (traction curve).
– Is COQUE_3D an appropriate choice for this problem, or would you recommend a different shell modelization?
– When I specify EPAIS in CARA_ELEM, how exactly is the pipe thickness taken into account for COQUE_3D? (Is it treated as a midsurface shell with this thickness, or in a different way?)
Large deformation and convergence issues (GROT_GDEP vs ELAS)
I expect pipe deformation larger than about 10% of the pipe thickness/diameter, so I tried using:
DEFI_MATERIAU(...)
...
COMPOR = _F(RELATION='VMIS_ISOT_TRAC', DEFORMATION='GROT_GDEP', ...)
However, with GROT_GDEP the STAT_NON_LINE analysis does not converge at all. If I switch to a purely elastic behaviour (ELAS in COMPORTEMENT), sometimes the computation converges.
– For this kind of indentation problem with significant plasticity and large deformation, is GROT_GDEP the right choice, or is there a better recommendation for shell elements?
– Are there any typical settings (time stepping, Newton parameters, etc.) that are usually needed to make COQUE_3D + plasticity + large deformation converge more robustly?
Contact definition and method (speed vs robustness)
I define contact between the torus indenter and the pipe surface, but I am not fully confident my contact definition is correct.
– Does my current DEFI_CONTACT setup look reasonable for shell–solid (or shell–rigid) contact in this case?
– Would you recommend the penalization method here, especially for large deformations, or is a different contact formulation preferable? I would also like to keep the computation time reasonably fast.
I am attaching my mesh files and .comm file for reference. Any hints, corrections, or best-practice examples for this kind of shell pipe indentation problem would be very much appreciated.
Thank you in advance for your help!

Thanks
Md Saleh Akram

stage-1.txt

5kB

Saleh

I could not upload the file as zip file. here is the link to access the mesh .med file. Any suggestion on modifying the mesh is appreciated [https://mega.nz/folder/ntQnEAIb#UKRk543pD4ENgS49uKBNaA]

I would like to add, simulation with small displacement load tends to converge with PETIT option. Whenever bigger displacement is given sometimes the model does not show any plasticity or the model does not converge.
So I switched to GROT_GDEP with COQUE_3D for bigger displacement load, for this the solver cannot solve no matter how many time-step I give.

I wanted to give internal pressure to the pipe. Is PRES_REP ok for my case?

Saleh

I am just adding this one does converge but I am afraid this would not give accurate result as 'PETIT' is only for small displacement and small rotation.

DEBUT(LANG='EN')

imesh = LIRE_MAILLAGE(identifier='0:1',
UNITE=20)

pmesh = LIRE_MAILLAGE(identifier='1:1',
UNITE=2)

pmesh_q9 = CREA_MAILLAGE(identifier='2:1',
MAILLAGE=pmesh,
MODI_MAILLE=_F(GROUP_MA=('pipe2d', ),
OPTION='QUAD8_9'))

pmesh_q9 = MODI_MAILLAGE(identifier='3:1',
reuse=pmesh_q9,
MAILLAGE=pmesh_q9,
ORIE_PEAU_3D=_F(GROUP_MA=('pipe2d', )))

mesh = ASSE_MAILLAGE(identifier='4:1',
MAILLAGE_1=imesh,
MAILLAGE_2=pmesh_q9,
OPERATION='SUPERPOSE')

model = AFFE_MODELE(identifier='5:1',
AFFE=(_F(GROUP_MA=('indtsolid', ),
MODELISATION=('3D', ),
PHENOMENE='MECANIQUE'),
_F(GROUP_MA=('pipe2d', ),
MODELISATION=('COQUE_3D', ),
PHENOMENE='MECANIQUE')),
MAILLAGE=mesh)

elemprop = AFFE_CARA_ELEM(identifier='6:1',
COQUE=_F(EPAIS=2.5,
GROUP_MA=('pipe2d', )),
MODELE=model)

rigid = DEFI_MATERIAU(identifier='7:1',
ELAS=_F(E=2000000.0,
NU=0.3))

loadf = DEFI_FONCTION(identifier='8:1',
NOM_PARA='INST',
VALE=(0.0, 0.0, 1.0, 1.0, 2.0, 1.0, 3.0, 1.0, 4.0, 0.0, 5.0, 0.0))

hardenf = DEFI_FONCTION(identifier='9:1',
ABSCISSE=(0.001, 0.024861075, 0.05156642, 0.090246491, 0.164044152, 0.32409948, 0.642396732, 1.249059062, 2.28380089, 4.004136),
INTERPOL=('LIN', ),
NOM_PARA='EPSI',
ORDONNEE=(21.0, 49.5991984, 99.19839679, 149.6993988, 199.2985972, 249.7995992, 299.3987976, 349.8997996, 399.498998, 450),
PROL_DROITE='LINEAIRE',
PROL_GAUCHE='LINEAIRE')

steel = DEFI_MATERIAU(identifier='10:1',
ELAS=F(E=210000.0,
NU=0.3),
TRACTION=F(SIGM=hardenf))

fieldmat = AFFE_MATERIAU(identifier='11:1',
AFFE=(_F(GROUP_MA=('pipe2d', ),
MATER=(steel, )),
_F(GROUP_MA=('indtsolid', ),
MATER=(rigid, ))),
MAILLAGE=mesh,
MODELE=model)

listr = DEFI_LIST_REEL(identifier='12:1',
DEBUT=0.0,
INTERVALLE=(_F(JUSQU_A=1.0,
NOMBRE=20),
_F(JUSQU_A=2.0,
NOMBRE=20),
_F(JUSQU_A=3.0,
NOMBRE=20),
_F(JUSQU_A=4.0,
NOMBRE=20),
_F(JUSQU_A=5.0,
NOMBRE=20)))

times = DEFI_LIST_INST(identifier='13:1',
DEFI_LIST=F(LIST_INST=listr),
ECHEC=F(ACTION='DECOUPE'),
METHODE='AUTO')

presf = DEFI_FONCTION(identifier='14:1',
NOM_PARA='INST',
VALE=(0.0, 0.0, 1.0, 0.0, 2.0, 1.0, 3.0, 0.0, 4.0, 0.0, 5.0, 1.0))

bc = AFFE_CHAR_MECA(identifier='15:1',
DDL_IMPO=_F(GROUP_MA=('pipefix', ),
LIAISON='ENCASTRE'),
MODELE=model)

contact = DEFI_CONTACT(identifier='16:1',
FORMULATION='DISCRETE',
LISSAGE='OUI',
MODELE=model,
ZONE=_F(GROUP_MA_ESCL=('pipe2d', ),
GROUP_MA_MAIT=('indtsurf', )))

indt = AFFE_CHAR_MECA(identifier='17:1',
DDL_IMPO=_F(DX=-14.0,
DY=0.0,
DZ=0.0,
GROUP_MA=('indtsolid', )),
MODELE=model)

pressure = AFFE_CHAR_MECA(identifier='18:1',
MODELE=model,
PRES_REP=_F(GROUP_MA=('pipe2d', ),
PRES=3.0))

resnonl = STAT_NON_LINE(identifier='19:1',
CARA_ELEM=elemprop,
CHAM_MATER=fieldmat,
COMPORTEMENT=(_F(DEFORMATION='PETIT',
GROUP_MA=('pipe2d', ),
RELATION='VMIS_ISOT_TRAC'),
F(DEFORMATION='PETIT',
GROUP_MA=('indtsolid', ),
RELATION='ELAS')),
CONTACT=contact,
CONVERGENCE=F(ITER_GLOB_MAXI=100,
RESI_GLOB_MAXI=1e-06),
EXCIT=(_F(CHARGE=bc),
_F(CHARGE=indt,
FONC_MULT=loadf),
F(CHARGE=pressure,
FONC_MULT=presf)),
INCREMENT=F(LIST_INST=listr),
MODELE=model,
NEWTON=_F(MATRICE='TANGENTE',
PREDICTION='TANGENTE',
REAC_ITER=1))

IMPR_RESU(identifier='20:1',
FORMAT='MED',
RESU=_F(RESULTAT=resnonl),
UNITE=80)

FIN()