Analysis of a Wall with 3D and 2D Meshes

Humberto · 2015-06-14 21:07:36

Hello everyone, I am doing an Analysis of a typical wall system with 3D elements on some elements of the Wall, and 2D (DKT) on the Steel studs. I am creating a
relationship between the 3D element and the 2D by the command LIAISON MAIL.

This should be pretty straight forward, but for a reason that I don't understand (yet) I get this error message:

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
! <S> Exception utilisateur levee mais pas interceptee. !
! Les bases sont fermees. !
! Type de l'exception : error !
! !
! les 8558 mailles imprimées ci-dessus n'appartiennent pas au modèle et pourtant !
! elles ont été affectées dans le mot-clé facteur : !
! PRES_REP !
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

I really don't understant why it says I have two meshes, and that they do not belong to the model.

Please help me understand the problem. The associated hdf and command.comm files are here:

https://drive.google.com/file/d/0B3df_K … sp=sharing (this is the Study2c.hdf)

And
https://drive.google.com/file/d/0B3df_K … sp=sharing (this is the command.com)

jeanpierreaubry · 2015-06-15 06:51:48

hello

it looks like group LoadY in not defined in AFFE_MODELE

jean pierre aubry

Humberto · 2015-06-15 17:23:53

jeanpierreaubry Thank you so much for allways replying and sharing your knowledge of Code_Aster.

I don't understand why should I include LoadY in AFFE_MODELE.

LoadY is just a group that includes a face where I want to apply a load.

but the volume that also includes the face LoadY is a 3D element grupo called MB which is already included in AFFE_MODELE.

AsterO'dactyle · 2015-06-15 19:05:15

Because you MUST have surface elements to apply pressure.
That's rule in Code_Aster

Humberto · 2015-06-15 23:11:02

Ok, thanks for pointing me that, I went with TOUT OUI on the AFFE_MODELE and the overloaded the rest of the model parts according to weather they were 3D or DKT

That solved the problem only to get me on with another error message, apparently the Noeud N15 I am trying to limit DRZ (I have nowhere in my command file DRZ=0 ) I tried changing my support condition to an edge group instead of a face, And got the same error message. Apparently the Node 15 belongs to my support face where I am applying the DDL in X Y and Z = 0

I think this node belongs both to a group where I am imposing a DDL to fix it, and might also be part of the LIAISON_MAIL But I am not sure if this should become problematic.

I am Totally lost, I will really appreciate your input once more and as allways, thanks for the previous input.

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
! <S> Exception utilisateur levee mais pas interceptee.
! Les bases sont fermees.
! Type de l'exception : error
!
! Erreur utilisateur:
! On cherche à imposer une condition aux limites sur le ddl DRZ
! du noeud N15.
! 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 ?
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

Last edited by Humberto (2015-06-15 23:18:30)

jeanpierreaubry · 2015-06-16 06:25:33

first of all it is very difficult to answer this question without the mesh
second it is not allowed if a node belong to a plate element to fix the rotation in an axis normal to the plane of the element

Humberto · 2015-06-16 17:00:19

jeanpierreaubry wrote:

It is very difficult to answer this question without the mesh

Sorry Jeanpierreaubry, I thought with the hdf file the mesh, geometry and groups were available, if not, I saved the mesh as a .med file that can be accessed here:

https://drive.google.com/file/d/0B3df_K … sp=sharing

jeanpierreaubry wrote:

it is not allowed if a node belong to a plate element to fix the rotation in an axis normal to the plane of the element

I am not limiting any rotations on any element, I have a group named Supp1 and I made a DDL_IMPO with only DX=0 DY=0 DZ=0. No DR's

I have some nodes that belong to the fixed support and also to the LIAISON_MAIIL groups (could that be a problem?) I really appreciate your help.

jeanpierreaubry · 2015-06-17 08:37:04

i had a quick look at your mesh
and i tried to run your problem with 12.2 version
looking at the .mess file i can see that there are :

1/
a very large number (over 4000) of double elements, unless you rally want that, this may induce completely wrong results

2/
568 element involved in the LIAISON_MAIL command do not belong to the model

Humberto · 2015-06-17 19:17:08

1.- Thanks for looking at my mesh, I don't understant how can I still have double elements, (does that mean two mesh elements at the same space?) in order to get rid of the double elements , I redid the geometry and left small spaces between things that were in contact before, and I suppose that the LIAISON_MAIL will take care of that.

Apparently redoing the mesh without intersecting areas did not work. Could this problem be caused by different Groups that intersect themselves? I have big groups to assign materials and smaller ones for loadings, fixing and for liaison.

2.- How could there be elements that not belong to the model if I inserted in the AFFE_MODELE one AFFE _ TOUT=OUI ?

Apparently I am really close from being able to run the model, so far it runs and crashes at the last command of my command.comm file, this is the message:

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

I tried using the ANGL_REP (0,0) but that didn't work either. changed that to VECTEUR(1,1,1) and the whole process goes and does not terminate with an error message, but no results..... I dont really care about the local coordinates as I will be mostly examining the Von Misses and Rankin Stress criteria.

I will really appreciate your help. and will post the resulting working model for the community to study / Use

The new command.comm is here:

https://drive.google.com/file/d/0B3df_K … sp=sharing

And the new mesh is here:

https://drive.google.com/file/d/0B3df_K … sp=sharing

For the mesh groups: here is the meaning:

Supp1 .... .........................Support grups to fix the model
LoadY, LoadX, LoadZ .... Self explanatory
M and S.......................... Master and Slave faces for LIAISON_MAIL to make the interaction between the face board and the steel studs of the wall.
Steel .............................. For assigning Materials
Magboard........................For assigning Materials
Foam...............................For assigning Materials

A picture of the model is this.
YvjyS8B

Last edited by Humberto (2015-06-18 20:05:09)

Attachments:
wall.png, Size: 39.29 KiB, Downloads: 311

jeanpierreaubry · 2015-06-19 07:53:21

if you have in the same group an element lying in the xoy plane and an element lying in the xoz plane
a VECTEUR(0,0,1), will be normal to the first hence its a projection on this element does not define a vector the calculation stops with an error

you have to split group or change the vector

even if

I dont really care about the local coordinates as I will be mostly examining the Von Misses and Rankin Stress criteria.

Code_Aster cares for you as this vector is necessary to calculate von Mises anyway!!

Humberto · 2015-06-22 19:35:00

Thanks! jeanpierreaubry, Actually I tried with VECTEUR(1,1,1) As I knew there were no surfaces normal to that vector, And also on MODI_MAILLAGE I included the ORIE_NORM_COQUE.

Interestingly the process ends with no error message, but I am gettin no results.

On the other hand, I have been trying really hard to identify the double elements, With no luck, I have no Idea how I have come to have them. My original .Brep file does not have any surfaces doubled. It Does have some surfaces parallel and very close one to the other, about 0.2mm but not in contact

Maybe the system is rounding their position and thus putting them on the same place? Could that be the case?

Thanks in advance.

The command.comm file with these latest changes:

https://drive.google.com/file/d/0B3df_K … sp=sharing

Last edited by Humberto (2015-06-22 19:45:00)

Humberto · 2015-06-24 00:04:30

Anybody?

#1 2015-06-14 21:07:36

Analysis of a Wall with 3D and 2D Meshes

#2 2015-06-15 06:51:48

Re: Analysis of a Wall with 3D and 2D Meshes

#3 2015-06-15 17:23:53

Re: Analysis of a Wall with 3D and 2D Meshes

#4 2015-06-15 19:05:15

Re: Analysis of a Wall with 3D and 2D Meshes

#5 2015-06-15 23:11:02

Re: Analysis of a Wall with 3D and 2D Meshes

#6 2015-06-16 06:25:33

Re: Analysis of a Wall with 3D and 2D Meshes

#7 2015-06-16 17:00:19

Re: Analysis of a Wall with 3D and 2D Meshes

#8 2015-06-17 08:37:04

Re: Analysis of a Wall with 3D and 2D Meshes

#9 2015-06-17 19:17:08

Re: Analysis of a Wall with 3D and 2D Meshes

#10 2015-06-19 07:53:21

Re: Analysis of a Wall with 3D and 2D Meshes

#11 2015-06-22 19:35:00

Re: Analysis of a Wall with 3D and 2D Meshes

#12 2015-06-24 00:04:30

Re: Analysis of a Wall with 3D and 2D Meshes

Board footer