Welcome to the forums. Please post in English or French.

You are not logged in.

#1 Code_Aster usage » Sigm_elno » 2013-08-07 17:48:50

Theobalt
Replies: 1

Hi!

In order to compute SIGM_NOEU, I need the field SIGM_ELNO. But, how do I get it? It is not a possible choice in CALC_ELEM. I don't know where else I should look.

I want to get stresses (tensor) at nodes. Maybe there is a better way?

I am with version 10.5.

Thank you!

#2 Re: Code_Aster usage » Composites in a non-linear model » 2013-06-18 18:52:28

Thomas DE SOZA wrote:

This would require developing a tangent matrix (the elastic stiffness matrix is OK for linear elasticity) and making sure only 'ELAS' behaviour is allowed.

Here is what I used. From what I understand, I use a tangent matrix and the ELAS behaviour is allowed. What am I doing wrong?

stat=STAT_NON_LINE(MODELE=model,
                   CHAM_MATER=Material,
                   CARA_ELEM=CELEM,
                   EXCIT=(_F(CHARGE=hold,FONC_MULT=LoadRamp,),
                         ),
                   CONTACT=con,
                   COMP_INCR=_F(RELATION='ELAS',DEFORMATION='PETIT',),
                   INCREMENT=_F(LIST_INST=incrm,),
                   NEWTON=_F(MATRICE='TANGENTE',REAC_ITER=1,),
                   CONVERGENCE=_F(RESI_GLOB_RELA=1e-4,ITER_GLOB_MAXI=25,),
                   SOLVEUR=_F(METHODE='MULT_FRONT',STOP_SINGULIER='NON',NPREC=20,),
                   ARCHIVAGE=_F(LIST_INST=timestep,),);

I also tried to run the model without the contact (in that case I block all DOF of the nodes involved in the contact).

Thank you for your help.

#3 Code_Aster usage » Composites in a non-linear model » 2013-03-17 19:52:44

Theobalt
Replies: 3

Hi!

I want to introduce a composite material in a non-linear model (there is contact elsewhere, but not with the composite body). Code Aster tells me that: "Les matériaux de coque homogénéisées (ELAS_COQUE ou DEFI_COQU_MULT) sont interdits en non-linéaire." What should I use then?

Thank you very much!

#4 Re: Code_Aster usage » Stresses from COQUE_3D and DKT elements » 2013-02-18 13:55:22

Bonjour!

Je crois bien que vous avez raison! Mais, il y a moyen d'obtenir des contraintes de flexion avec des éléments DKT?

For those that do not speak French (nobody is perfect they say... smile ), how can we get flexural stresses with DKT elements?

Merci!

#5 Re: Code_Aster usage » Stresses from COQUE_3D and DKT elements » 2013-02-18 02:09:00

I paste here the code following STAT_NON_LINE

RESU=CALC_ELEM(reuse =RESU,
                 RESULTAT=RESU,
                 TYPE_OPTION='SIGM_STRUCT',
                 OPTION='SICO_ELNO',);

RESU=CALC_ELEM(reuse =RESU,
                 RESULTAT=RESU,
                 TYPE_OPTION='CRIT',
                 OPTION='SIEQ_ELNO',);

RESU=CALC_ELEM(reuse =RESU,
                 RESULTAT=RESU,
                 TYPE_OPTION='EPSI',
                 OPTION='EPSI_ELNO',);

RESU=CALC_NO(reuse =RESU,
               RESULTAT=RESU,
               OPTION=('SIEQ_NOEU','REAC_NODA','EPSI_NOEU',),);

With COQUE_3D, I project the data on a TRIA6 mesh:

Postpro=AFFE_MODELE(MAILLAGE=MAIL_Q,
                    AFFE=_F(TOUT='OUI',
                            PHENOMENE='MECANIQUE',
                            MODELISATION='3D',),);

PROJ=PROJ_CHAMP(RESULTAT=RESU,
                MODELE_1=MODE,
                MODELE_2=Postpro,);

IMPR_RESU(MODELE=Postpro,
          FORMAT='MED',
          RESU=_F(MAILLAGE=MAIL_Q,
                  RESULTAT=PROJ,
                  TOUT='OUI',),);

I hope someone sees where the problem is.
Thank you very much!

#6 Code_Aster usage » Stresses from COQUE_3D and DKT elements » 2013-02-17 20:59:22

Theobalt
Replies: 5

Hi!

I am working on a simple test case: a plate under a constant pressure with four sides fixed (all DOF=0). I try it with both COQUE_3D and DKT elements, to test quadratic and linear elements. I basically use the same code, but for the postprocessing I get different issues.

1) with COQUE_3D elements, I can't get the stress tensor, but I have the Von Mises stresses;
2) with DKT elements, I can't get the Von Mises stresses, but I get the stress tensor (not at nodes, but I can get the maximum value on each element)

The point is that I would like to have both Von Mises stresses and the stress tensor at each node.

What is wrong with my script?

I use SALOME-MECA 2012.2, running on ubuntu 10.04 64 bits. I join the files.

Thank you very much!

#7 Re: Code_Aster usage » Problem with MASSIF_COQUE (LIAISON_MAIL) » 2012-04-25 13:40:04

Hi Jean-Pierre!

I managed to get it run!

By curiosity, how did you do it? With MASSIF, COQUE_MASSIF or MASSIF_COQUE? Did you also use VECTEUR in AFFE_CARA_ELEM to solve the issue with the normals (see error message reported above)?

The model also run with TETRA10 on the bracket.

COQUE_3D works when glued perpendicular on a surface. It seems that only the surface-to-surface gluing that causes problems.

#8 Re: Code_Aster usage » Problem with MASSIF_COQUE (LIAISON_MAIL) » 2012-04-24 17:53:07

Effectively, changing from COQUE_3D to DKT seems a good choice from what I can see rapidly in the documentation.

For the welding of the aluminium bracket on the steel tube, I will have to answer in French (I hope you don't mind!).

La pièce est simplement brasée sur le tube. Comme la force appliquée vient essentiellement plaquer la pièce en aluminium sur le tube, il y a relativement peu de cisaillement entre les deux pièces. En on parle de forces autour de 500N. Alors, pas besoin de technologies de l'espace ici. wink

Hence, for the moment, this is the general behaviour which interests us, not the behaviour of the welding, which we know will not fail.

#9 Re: Code_Aster usage » Problem with MASSIF_COQUE (LIAISON_MAIL) » 2012-04-24 16:44:59

Nice Jean-Pierre!

You simply changed to a linear mesh and everything worked? The problem was the use of quadratic elements?

I have been told that linear tetraedrons were too stiff and were giving wrong stresses in consequence. This is why I was using TETRA10 and TRIA7 elements. Should I understand that linear shells do not have this kind of problem?

Thank you very much for your help.

#10 Re: Code_Aster usage » Problem with MASSIF_COQUE (LIAISON_MAIL) » 2012-04-24 14:48:44

jeanpierreaubry wrote:

are you trying to study the weldment behavior ar the tube behavior?

The behavior of the tube.

#11 Re: Code_Aster usage » Problem with MASSIF_COQUE (LIAISON_MAIL) » 2012-04-24 11:44:57

jeanpierreaubry wrote:

first one detail, maybe you already know

But since GROUP_NO_MAIT does not exist

you can create groups of nodes with

DEFI_GROUP(
.................
  CREA_GROUP_NO=(
	_F(TOUT_GROUP_MA='OUI',), #for all groups
	_F(GROUP_MA='thisgroup',),#or for this group
	),
);

Yes, I am aware, my point was that the option GROUP_NO_MAIT does not exist with MASSIF_COQUE.

jeanpierreaubry wrote:

the last med file is unusable without the comm file that goes with it

It should work with the .comm previously posted. I join my current .comm, but it is quite messy as I tried a lot of things to get it to work...

jeanpierreaubry wrote:

the more generaly
you seem to struggle with LIAISON_MAIl in different directions,
looking at your model i do not understand what you try to simulate
this would be interesting to know to help you solve the problem

for example i do not understand why yo use TRIA 6 and  TETRA10 element at first

This is about all my client permits me to put on the forum. What I can say is that things like that bracket are welded on the tube, and basically forces will push the bracket on the tube. I use quadratic elements since they generally give better results than linear ones, especially with tetraedrons. Do not hesitate to ask further questions if needed, I will do my best to answer as soon as possible.

#12 Re: Code_Aster usage » Problem with MASSIF_COQUE (LIAISON_MAIL) » 2012-04-24 00:31:08

I got back to COQUE_MASSIF with GROUP_NO_ESCL with a mesh that generates less NOEUDS LAGRANGE, so I have enough ressources on my computer to get something. I joined the mesh to the post. But, that something is yet another error...

   !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
   ! <S> Exception utilisateur levee mais pas interceptee. !
   ! Les bases sont fermees.                               !
   ! Type de l'exception : error                           !
   !                                                       !
   !  l'axe de rfrence est normal  un lment de plaque !
   !  calcul option impossible                             !
   !  orienter ces mailles                                 !
   !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

Will I ever see the end of that problem?!?

Is there a way to get rid of these center nodes the time to create that welding?

Alternatively, I was thinking to impose a contact with "infinite" friction (coefficient of friction of let's say 1000) to create that weld, but I will need frictionless contact on another part of the model. From what I understand STAT_NON_LINE can only take one type of contact (but that can be applied to many places), hence this is not a solution for me. Do I understand well?

Thanks again.

#13 Re: Code_Aster usage » Problem with MASSIF_COQUE (LIAISON_MAIL) » 2012-04-23 23:02:58

keeswouters wrote:

cannot you use TYPE_RACCORD='MASSIF',  iso COQUE_MASSIF?
I would say you donot need to restrict the rotational DOFs of the shell elements, just the translational DOFs would be just right in your case.

Of course, just restricting the translational DOFs is enough, but with that method I get issues with the center nodes not having translational DOFs...

#14 Re: Code_Aster usage » Problem with MASSIF_COQUE (LIAISON_MAIL) » 2012-04-22 21:09:35

Effectively, I think Richard was right. But since GROUP_NO_MAIT does not exist, I can't see how I can use MASSIF_COQUE without that problem with center nodes.

COQUE_MASSIF with GROUP_MA_ESCL leads to an erreur de programmation. So that path is doomed...

So I tried COQUE_MASSIF with GROUP_NO_ESCL and the connection seems to be done correctly, but I cannot verify it because the problem has 225598 nodes and demands too much memory to end without error. This is very confusing since there are only about 42000 nodes in the mesh (+ the middle nodes for the about 14000 TRIA7 elements, so about 56000 nodes in the end). The 170810  NOEUDS "LAGRANGE" are quite surprising to me. Where do they come from? Were they generated by LIAISON_MAIL?

So, after hours of work, I am still unable to weld a bracket on a tube meshed with shell elements and get simulation results... and this is only a small part of the overall model...

Alternatively, do the tube meshed with only one volumetric element along the thickness has any chance of giving good results? More than one element along the thickness would make the model too big to run on my computer, so that is not an option at all.

Thanks again.

#15 Re: Code_Aster usage » Problem with MASSIF_COQUE (LIAISON_MAIL) » 2012-04-21 23:27:22

Hi Kees!

Ok, maybe I am blind, but I really can't see where I am doing something different than in your example KW_V04_massive_coque_coque3d (which runs perfectly on my computer, with version STA10.3). I even rewrote the script to place the commands in the exact same order (although not convince it could change something wink ), and I still get that strange error message about nodes not having the DX dof (how is it possible?!?).

What am I missing?

Again, I join the files.

#16 Code_Aster usage » Problem with MASSIF_COQUE (LIAISON_MAIL) » 2012-04-21 16:49:29

Theobalt
Replies: 20

Hi!

In short, I am trying to weld a bracket (3D) on a tube (COQUE_3D). When it comes to LIAISON_MAIL, I get this message:

   
   !-----------------------------------------------------------------------!
   ! <EXCEPTION> <MODELISA8_30>                                            !
   !                                                                       !
   ! Erreur utilisateur:                                                   !
   !     On cherche à imposer une condition aux limites sur le ddl DX      !
   !     du noeud NN9722.                                                  !
   !     Mais ce noeud ne porte pas ce ddl.                                !
   !                                                                       !
   !     Conseils :                                                        !
   !      - vérifier le modèle et les conditions aux limites :             !
   !         - le noeud incriminé fait-il partie du modèle ?               !
   !         - le noeud porte-t-il le ddl que l'on cherche à contraindre ? !
   !-----------------------------------------------------------------------!
   
What can I do with that?!?

Here is the .comm:

DEBUT();

##LECTURE DU MAILLAGE
MAIL_BRA=LIRE_MAILLAGE(UNITE=20,
                     FORMAT='MED',);
MAIL_TUB=LIRE_MAILLAGE(UNITE=26,
                     FORMAT='MED',);

MAIL=ASSE_MAILLAGE(MAILLAGE_1=MAIL_BRA,
                   MAILLAGE_2=MAIL_TUB,
                   OPERATION='SUPERPOSE',);

MAILMODI=CREA_MAILLAGE(MAILLAGE=MAIL,
                       MODI_MAILLE=(_F(PREF_NOEUD='NN',GROUP_MA='tubecoq1',OPTION=('TRIA6_7'),),),);

MAILMODI=MODI_MAILLAGE(reuse =MAILMODI,
                   MAILLAGE=MAILMODI,
                   ORIE_PEAU_3D=_F(GROUP_MA=('BraSouMA',),),);

##MODELE
model=AFFE_MODELE(MAILLAGE=MAILMODI,
                  AFFE=(
                        _F(TOUT='OUI',
                           PHENOMENE='MECANIQUE',
                           MODELISATION='3D',),
                        _F(GROUP_MA='tubecoq1',
			   PHENOMENE='MECANIQUE',
			   MODELISATION='COQUE_3D',),
                       ),
                 );
CARA=AFFE_CARA_ELEM(MODELE=model,
                    COQUE=_F(GROUP_MA='tubecoq1',EPAIS=0.0005,),
                   );

##MATERIAUX
ACIER=DEFI_MATERIAU(ELAS=_F(E=210.0e9,
                          NU=0.27,),);
ALU=DEFI_MATERIAU(ELAS=_F(E=70.0e9,
                          NU=0.33,),);

Material=AFFE_MATERIAU(MAILLAGE=MAIL,
                       MODELE=model,
                       AFFE=(_F(GROUP_MA='brackeMA',
                                MATER=ALU,),
                             _F(GROUP_MA=('tubecoq1',),
                                MATER=ACIER,),
                             ),
                      );

##CONDITIONS LIMITES
hold=AFFE_CHAR_MECA(MODELE=model,
                    LIAISON_MAIL=(
                                 _F(GROUP_MA_MAIT='tubecoq1',
                                    GROUP_MA_ESCL='BraSouMA',
                                    TYPE_RACCORD='MASSIF_COQUE',
#                                    DDL_MAIT='DNOR',DDL_ESCL='DNOR',
                                   ),
                                 ),
                   );

##RESULTATS
IMPR_RESU(FORMAT='MED',
          RESU=_F(MAILLAGE=MAIL,),);

FIN();

I join the files, so you could have a look at it and even try it at home or at the office. It contains only the commands to test LIAISON_MAIL. Of course, there will be more in the future! wink

Thanks a lot!

#17 Re: Code_Aster usage » Modeling an hinge in a 3D model » 2012-03-31 14:40:55

Hi again!

One thing that I do not understand in your tutorial: how do you set that surfaces are glued together or can slide on each other? I do not understand how you introduce this difference in your .comm. I get the impression that with LIAISON_MAIL surfaces are always glued together, hence no free rotation would be allowed. Am I wrong?

Thanks!

#18 Code_Aster usage » Easy question: how to set a force on a surface? » 2012-03-23 23:20:05

Theobalt
Replies: 2

Hi!

From what I understand FORCE_FACE requires a force per unit area. Is there a way to say to Code_aster to set a let's say 1000N force on such or such area, or do I have to manually compute the area of the given surface, compute the resulting force per unit area and then enter that value at the FORCE_FACE keyword?

In short, with Abaqus, Cosmos, and others, we simply indicate for example: "put 1000N on the surface A." Is there an equivalent with Code_Aster?

Thanks!

#19 Re: Code_Aster usage » Modeling an hinge in a 3D model » 2012-02-18 01:07:51

Claws wrote:

That's what I mean by the above smile

Hum... I am a bit confused... Sorry for being so blind!

#20 Re: Code_Aster usage » Modeling an hinge in a 3D model » 2012-02-17 03:31:04

It looks very useful!!! Does it work also if we model the pin as a beam, to save further computation time?

#21 Code_Aster usage » Modeling an hinge in a 3D model » 2012-02-16 01:45:09

Theobalt
Replies: 6

Hi!

Someone here ever modeled an hinge in a 3D model? The bracket and the bar are modeled with 3D elements, and I was wondering what would be the most efficient way to model the pin. The lazy way would be to model a 3D pin with frictionless contacts with the bracket and the bar, but it would most likely need a relatively dense mesh to correctly solve the contact problem, which quickly becomes quite time consuming.

Someone knows a more elegant way to solve that problem?

Thank you very much.

#22 Re: Code_Aster usage » Python Code in *.comm - full Python possible? » 2011-07-21 13:47:42

Hi!

You should take a look at [U1.03.02] Méthodes Python d'accès aux objets Aster.

You can do whatever you want. For post-processing, I often extract values from Code_Aster results, treat them with python scripts, use the results as a new input for some Code_Aster functions, extract the new results, etc. One could for example write an optimisation script in python and use it in a loop with Code_Aster functions and modify your model at each loop until you get the wanted results.

Your creativity is about the only limit.

#23 Code_Aster usage » Restart a crashed computation » 2011-07-21 13:40:41

Theobalt
Replies: 1

Hi!

I have a few very long computations that crashed because of a lack of CPU time. How can I restart them from the point where it crashed? Is it sufficient to point to the saved base in the .export and use the POURSUITE keyword in the .comm (along with commenting lines that define concepts which are already defined in the saved base)?

Also, may I suggest to make the CPU max time infinite by default? If it is taking too much time to the opinion of the user, it is always time to kill the job.

Thanks!

#24 Re: Code_Aster usage » [SOLVED]Interpretation of .mess (AFFE_MODELE) - easy one ;) » 2011-07-20 15:52:00

Thomas DE SOZA wrote:

3D_SI must be put after 3D.

Ok! I thought the order was the order of priority, meaning 3D_SI is affected first, and then 3D the the remaining of the model which do not support 3D_SI.

#25 Code_Aster usage » [SOLVED]Interpretation of .mess (AFFE_MODELE) - easy one ;) » 2011-07-20 15:27:26

Theobalt
Replies: 4

Hi!

I have this in my .mess

   MODE=AFFE_MODELE(MAILLAGE=MAIL,
                   INFO=1,
                   AFFE=_F(PHENOMENE='MECANIQUE',
                           TOUT='OUI',
                           MODELISATION=('3D_SI','3D',)),
                   PARTITION=_F(PARALLELISME='CENTRALISE'),
                   VERI_JACOBIEN='OUI',
                   );


 SUR LES        34124 MAILLES DU MAILLAGE MAIL    
    ON A DEMANDE L'AFFECTATION DE        34124
    ON A PU EN AFFECTER                  34124

    MODELISATION      ELEMENT FINI      TYPE MAILLE          NOMBRE
                      MECA_ARETE3       SEG3                   1546
                      MECA_FACE6        TRIA6                  8166
    3D                MECA_TETRA10      TETRA10               24412

Should I understand that Aster computed using fully integrated quadratic tetrahedrons (3D) instead of elements with reduced integration (3D_SI)?

Otherwise, if i use only 3D_SI in MODELISATION, I get:

   MODE=AFFE_MODELE(MAILLAGE=MAIL,
                   INFO=1,
                   AFFE=_F(PHENOMENE='MECANIQUE',
                           TOUT='OUI',
                           MODELISATION='3D_SI'),
                   PARTITION=_F(PARALLELISME='CENTRALISE'),
                   VERI_JACOBIEN='OUI',
                   );


 SUR LES        70243 MAILLES DU MAILLAGE MAIL    
    ON A DEMANDE L'AFFECTATION DE        70243
    ON A PU EN AFFECTER                  68271

    MODELISATION      ELEMENT FINI      TYPE MAILLE          NOMBRE
                      MECA_FACE6        TRIA6                 15260
    3D_SI             MECA_TETRS10      TETRA10               53011

Code aster do not "affecte" (how do we say in English?) the model to all elements. Any possible bad consequences to that? For the moment, it is still computing without any alarm or error message.