No convergence with an Mfront law in code_aster

[Max-EDFLMT]

Hi,

I'm trying to use an MFront law in a code_aster model. it is adapted from the one sent in #130 (I added Voce hardening law to it).

It's working as expected in Mtest and on a small domain with periodic boundary conditions in code_aster.
When I try on my real structure, code_aster never success in integrating the law. It works with a native aster law on the same model or with Mfront FCC crystal plasticity law.

I tried the keyword @AsterGenerateMTestFileOnFailure but I don't know where I'm supposed to find the files in the computation repository.

Any hints on how to debug this ?

Thanks in advance,

Maxime

OrthotropicHosford_Voce.txt

[thelfer]

Hi @Max-EDFLMT.

Good to see that you made progresses. Very interested to see what will be the final result

First of all, have you read the last part of this tutorial ?
You probably shall use the Jacobi eigen solver and either Algorithm 3 or Algorithm 4.

Concerning code_aster and the MTest file generation, I don't remember the details, but the basic idea is to put the call the STAT_NON_LINE in try/except bloc and do something like this:

try:
    STAT_NON_LINE(...)
except:
    os.system("cp *.mtest <insert_directory_where_to_copy_the_mtest_files>")

[Max-EDFLMT]

Thank you anyway. I will ask on code_aster forum this monday !

[thelfer]

My job !

[Max-EDFLMT]

It turned out that there is a keyword GROUP_MA in COMPORTEMENT = _F(...) to mention the element group you want to assign the previously mentionned DEFI_MATERIAU. I did not know it and thought everything was to be declared before.
Right now, everything seems to work !

[thelfer]

Great !

[rian44s]

Hi both,

I am a bit late on this discussion but I have, I think a very similar problem.
I have a simulation with multiple grains running perfectly in code-aster with a meric-cailletaud law, small strain.
I have downloaded MericCailletaudSingleCrystalViscoPlasticity.mfront to compare.
Then the only modification I have made in aster python launcher file are:

• monMateriau = DEFI_MATERIAU( ELAS =_F(E = young,NU = poisson),
  MFRONT =_F( LISTE_COEF = (young,young,young,poisson,poisson,poisson,cissail,cissail,cissail,
  N,K,tau0,Q,B,D,C,H0,H1,H2,H3,H4,H5,H6 )

  instead of

• monMateriau = DEFI_MATERIAU(ELAS_ORTH=_F(E_L= young,E_T= young,E_N= young,G_LT= cissail,G_TN= cissail,G_LN= cissail,NU_LT= poisson,NU_TN= poisson, NU_LN= poisson,),
  MONO_VISC1=_F(N=N, K=K,C=C),
  MONO_ISOT1=_F(R_0=Tau0, Q=Q,B=B,H1=H0,H2=H1,H3=H4,H4=H5,H5=H6),
  MONO_CINE1=_F(D=D))

then

• DEFI_FICHIER(FICHIER=path_mfront,UNITE=33)
  MonBehaviour = CREA_LIB_MFRONT(NOM_COMPOR=mfront_name,UNITE_MFRONT=33,)

  instead of

• MonBehaviour = DEFI_COMPOR(MONOCRISTAL =(_F(MATER = monMateriau,ELAS ='ELAS_ORTH',ECOULEMENT ='MONO_VISC1',ECRO_ISOT = 'MONO_ISOT1',ECRO_CINE = 'MONO_CINE1',FAMI_SYST_GLIS = 'OCTAEDRIQUE')))

and in Stat_non_lin

• COMPORTEMENT =_F( RELATION = 'MFRONT',
  COMPOR_MFRONT = MonBehaviour,
  DEFORMATION = 'PETIT',
  ITER_INTE_MAXI = 1000,
  RESI_INTE_MAXI =1.0E-12,
  ),

instead of

• COMPORTEMENT=_F(COMPOR=MonBehaviour,DEFORMATION='PETIT',RELATION='MONOCRISTAL'),

I eventually get this error:
<MECANONLINE9_3>
Arrêt suite à l'échec de l'intégration de la loi de comportement.

@Max-EDFLMT you said that adding GROUP_MA in comportement helped but I don't understand, the default value is 'all' and link between behavior and domain is already done in
fieldmat = AFFE_MATERIAU(AFFE=_F(MATER=(monMateriau, ),TOUT='OUI'),MODELE=monModel), isn't it ?

I would be happy to see what worked for you.

Thank you in advance,
Rian SEGHIR