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

You are not logged in.

#1 Re: Code_Aster usage » problems with the bi-linear leaf spring example » 2022-12-08 15:14:39

Dear Jean Pierre,

inspired by your picture i modified the first time the parameter "Step" inside the view options of gmsh.

later then i found out that also ParaVis has in its menu the parameter "Time", which i had never used before.

Sorry for all my useless posts.

Now i will work through all my models again.


#2 Re: Code_Aster usage » problems with the bi-linear leaf spring example » 2022-12-07 17:50:01


here my last results.

The run commes to a good end (EXECUTION_CODE_ASTER_EXIT_8365=0), but all fields in the rmed-file are zero?!

The rmed-file seems to be to big as an attachement.


#3 Re: Code_Aster usage » problems with friction and STAT_NON_LINE » 2022-12-07 15:00:47

Hello Mario,

I already knew that with nonlinear models I need some kind of quasi-time to calculate them. It is not about a real time, it is not a dynamic model.

I want to apply the lateral force on box2 (side2) only to see how it transfers to box1 via friction. So there should be explicitly no displacement. For a displacement I would need a mass at least for box2.


#5 Re: Code_Aster usage » problems with friction and STAT_NON_LINE » 2022-12-07 14:30:18

Dear Jean Pierre,

I could not get my system to work with your hints, so I rebuilt this simple system hoping to learn something from it.

I am now retired, but I worked professionally with FEMAP a few years ago. It was also about problems with friction, but mainly about material behavior between elastic and plastic. Is it possible that from this previous experience I make mistakes when working with code_aster?


#6 Re: Code_Aster usage » problems with friction and STAT_NON_LINE » 2022-12-06 18:24:54


I just wanted to calculate a simple system with static friction, i.e. without displacement. Therefore the displacement force (side2) is also smaller than the normal force (load2) x coulomb.

Of course I got the error message <EXCEPTION> <FACTOR_11> several times. There it is referred to problems with boundary conditions or rigid body motions. That the object should not be held with the contact, I had read also already in a forum contribution. But how should I hold Box2 without contact? Then the normal force would also lose its effect on the contact surface.

I had also tried to include the adhesive connection in the calculation and to remove it against the displacement force with unload=DEFI_FONCTION(NOM_PARA='INST',VALE=(0.0, 1.0, 1.0, 0.0)). The calculation then runs through, but the result is meaningless.

I have now set COEF_PENA_CONT and COEF_PENA_FROT to 2,100,000. Unfortunately this does not help.
How do I leave DEFI_CONTACT in the calculation and exclude the friction? With FROTTEMENT='SANS'? Do I have to change something in STAT_NON_LINE then?

Master and slave are MAIT and ESCL respectively target and follower? Box2 should be master, Box1 slave?

By the way, I work with Salome Meca from the file SM-2021-w64-0.4.zip under Windows or with salome_meca-lgpl-2021.1.0-2-20220817-scibian-9.sif under Ubuntu 22.04. Could there be differences to using code_aster from the command line? I have not yet been able to bring myself to try this?.

Attached is another attempt (comm file and output) that calculates to completion, but the results are nonsensical. All values in the rmed file are zero. Whereas the scale in ParsVis always goes up to 1.2e-38, but when viewed in gmsh it goes up to 264 (SIGM,SIEF) or 295 (DEPL) only for SIEQ is zero?!


#7 Code_Aster usage » problems with friction and STAT_NON_LINE » 2022-12-05 15:25:03

Replies: 9


I still need some help. In the meantime, I've been tweaking all sorts of parameters for a few weeks.
Calculations with MECA_STATIQUE I get. But when friction and STAT_NON_LINE come into play, I get nowhere. Something is missing, or there is something too much!

Here is a simple example. A box1 (200x200x100) is fixed to the base (fix1). A Box2 (100x100x100) is centered on Box1 (Boxes_glued_Mesh.jpg). A force FORCE_FACE=_F(FZ=-100.0, GROUP_MA=('Top_2', )) pushes Box2 onto Box1. A second force FORCE_FACE=_F(FY=5.0, GROUP_MA=('Side_2', )) pushes Box2 in y-direction. For MECA_STATIQUE both boxes are glued together with LIAISON_MAIL (Boxes_glued_DEPL.jpg).

For STAT_NON_LINE there should be friction (Coulomb=0.1) between the boxes.

Many parameters for DEFI_CONTACT and STAT_NON_LINE I have gathered from various posts in this forum.

So: Where do I do something wrong?


#8 Re: Code_Aster usage » problems with the bi-linear leaf spring example » 2022-11-23 15:33:52

Thank you,

I have realized that I still have to learn a lot about gmsh ;-)

attached for comparison the results of case 3 (static) and case 4 (dynamic) displayed with ParaVis and gmsh

and comm and export file from case 4

the run is made with salome_meca-lgpl-2021.1.1-2-20220817-scibian-9.sif (Salome 9.7.0)
on singularity-ce_3.10.4-jummy-amd64.deb
on go 1.18.8 comming with Ubuntu 22.04.1 LTS

the gmsh package comming with Ubunto (version 4.8.4) comms without support vor MED files
therefore i use gmsh version 4.11.0 downloaded vrom gmsh.info

Regards Andreas

#9 Re: Code_Aster usage » problems with the bi-linear leaf spring example » 2022-11-22 15:57:35


there is no result looking like Figure 13.5 (page 176)?!

ParaViS show a result_DEPL_Magnitude from 0 to 1.2e-38 but part1 and part2 are all deep blue.
gmsh shows a result_DEPL out of the same file from 0 to 1.16 but no colored scale on the parts.

Regards Andreas

#10 Re: Code_Aster usage » problems with the bi-linear leaf spring example » 2022-11-22 10:51:45


attached is an output from a case 4 run with reduced complexity

I am not able to find a reference to a solution, to get a reasonable result

Regards Andreas

#11 Re: Code_Aster usage » problems with the bi-linear leaf spring example » 2022-11-15 20:16:10


under salome_meca-lgpl-2021.1.0-2-20220817-scibian-9.sif
ther is although only ORIE_PEAU (no ORIE_PEAU_2D nor ORIE_PEAU_2D)

attached the output under Linux

my computer:
Core i5-8400 CPU (2.80GHz, 6 Kerne)
48 GB RAM (2400 MHz)
M.2 SSD (500GB, 970 EVO PLUS)
PNY 5GB D5 X Quadro P2200

i am able to run a linux on a comman line, but i am not familiar with python,
therefore I prefer AsterStudy
to use code_aster with the command line will be possible, but least me somme days of trying

Regards Andreas

#12 Re: Code_Aster usage » problems with the bi-linear leaf spring example » 2022-11-15 10:44:25


thanks for your answer

AsterStudy gave me no ORIE_PEAU_2D and no ORIE_PEAU_3D, ownly ORIE_PEAU is possible

Inside u4.23.04.pdf i found:
This keyword serT to reorientate meshs of edge so that their normals are coherent (towards the
outside of the matter). This keyword is valid for a two-dimensional study (2D) or three-
dimensional (3D). It is an essential precondition if, for example, one wants to apply a loading of
pressure to this “skin”.

it seem to me, that this is not the problem

therefore my output (ownly for case 4) attached

Regards Andreas

#14 Code_Aster usage » problems with the bi-linear leaf spring example » 2022-11-13 09:59:49

Replies: 16

I have tried to reproduce the example from "Beginning with code_aster" by Jean-Pierre Aubry starting on page 145.
I work with Salome-Meca for Windows (SM-2021-w64-0.4.zip from code-aster-windows.com) which is based on Code_Aster for Windows (code-aster_v2021_std.msi also from code-aster-windows.com). But the result is the same with salome_meca-lgpl-2021.1.0-2-20220817-scibian-9.sif under Ubuntu 22.04.1 LTS.
1st problem: From my CAD program I could not create part1 and part2, which would have contained the points load1p and move2p, as well as a second volume (pin1v next to part1v) after export and import via STEP format. Even typing the .geo-files did not help; Gmsh version 4.10.5 reported errors when reading them.
So I had to rebuild the geometry of part1 and part2 with Gmsh. I attach here the .geo files. The .med files can be exported from Gmsh and then imported into the mesh module of Salome-Meca.
2nd problem: I did not succeed to realize the iteration loop for case 1 to case 5 in Salome-Meca. Therefore, in the attached .comm file, there are result_1 to result_5 under Analysis.
3rd problem: For the static analyses (result_1 to result_3) there are nice results (see attachments Figure 13_2_low.jpg, Figure 13_3_low.jpg and Figure 13_4_low.jpg).
Unfortunately, the dynamic analyses (STAT_NON_LINE) do not provide meaningful results. Where is the error?
Regards Andreas

#15 Code_Aster usage » Use of CALC_CHAMP options » 2022-07-25 15:27:24

Replies: 1

I have worked professionally until 2020 for simple static FEM with SolidEdge simulation (NX Nastran 11.0) and for more complex with FEMAP. Now I am retired and started to learn Salome Meca (Salome-Meca 2021 for Windows (SALOME 9.7.0) on Windows 10, salome_meca 2021.0.0-2-20211014 Scibian9 singularity container on Ubuntu).
My first project was a bending beam made of steel (E=210000 N/mm^2, Poisson=0.3), fixedly clamped on one side, cross section 100x100 mm, length 1000 mm. At the free end a force of 20000 N (FORCE_FACE: FZ=-2.0 N/mm^2).
Analytically, a max. deflection of 3.8096 mm and a max. bending stress of 120 N/mm^2 results.
With SolidEdge simulation I get 3.84 mm, a max. von Mises stress of 167 MPa and a max. normal stress in X-direction of 145 MPa.
With Salome Meca, I get 3.8 mm with a tetrahedral mesh, a max von Mises stress of +120/-130 N/mm^2 (SIEQ_NOEU and VMIS_SG) with a max magnitude of 360 N/mm^2.
In the output (.med file) I get additional fields by CALC_CHAMP only with
All options xxxx_NOEU.
Why gave xxxx_ELGA and xxxx_ELNO Options no further fields in output? Is there anything else to configure?