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

You are not logged in.

#1 2019-05-20 11:28:09

nirmaljoshi
Member
From: Japan
Registered: 2018-10-12
Posts: 181

[solved]help to find error

I have made two set of meshes (consisting 3D and BARRE) for an analysis with only minor differences. Now, the anlysis for one mesh runs without error while for another one, it give factorization error. I already used days to find the reason but in vain.

Can anyone please help. I have attachd both working mesh and error mesh.

Last edited by nirmaljoshi (2019-05-22 12:47:38)


Attachments:
upload.zip, Size: 12.23 KiB, Downloads: 76

Offline

#2 2019-05-20 12:24:27

manonB
Member
From: Allemagne
Registered: 2019-04-02
Posts: 68

Re: [solved]help to find error

Hello,

I'm running the working one to see. But the "not working" Folder is empty.

Manon

Offline

#3 2019-05-21 11:52:58

nirmaljoshi
Member
From: Japan
Registered: 2018-10-12
Posts: 181

Re: [solved]help to find error

Sorry for the mistake.
Here is the zip with both files.


Attachments:
upload.zip, Size: 26.28 KiB, Downloads: 84

Offline

#4 2019-05-21 15:23:52

jeanpierreaubry
Guru
From: nantes (france)
Registered: 2009-03-12
Posts: 3,638

Re: [solved]help to find error

hello

in working directory

i am not a specialist of concrete
but
with these conditions i am afraid the result you get is meaningless

CONVERGENCE=_F(
	ARRET='NON',
	ITER_GLOB_MAXI=200,
	RESI_GLOB_MAXI=1.0,
	RESI_GLOB_RELA=0.1,
),

i can notice that in the mess file at an early stage of the computation (time 8 / 150)
you have this warning

<Erreur> the maximum number of iterations of Newton is reached 

 <Action> One stops computation. 

jean pierre aubry


consider reading my book
freely available here https://framabook.org/beginning-with-code_aster/

Offline

#5 2019-05-22 05:05:03

nirmaljoshi
Member
From: Japan
Registered: 2018-10-12
Posts: 181

Re: [solved]help to find error

@jeanpierreaubry
Thank you for providing time to review the analysis files.
Currently, the issue is not the convergence. The issue is that the file in "not-working" folder do not run at all while another file with almost same configuration runs without problem. If we could settle the reason for this behaviour then we can later talk about the convergence issues. Hope for your kind help.

Offline

#6 2019-05-22 06:24:31

jeanpierreaubry
Guru
From: nantes (france)
Registered: 2009-03-12
Posts: 3,638

Re: [solved]help to find error

Currently, the issue is not the convergence.
.........
runs without problem

if you call convergence  or running without problem what i can see in the .mess file with the use of ARRET='NON' on time 8 on 150
we do not speak the same language


consider reading my book
freely available here https://framabook.org/beginning-with-code_aster/

Offline

#7 2019-05-22 07:06:22

nirmaljoshi
Member
From: Japan
Registered: 2018-10-12
Posts: 181

Re: [solved]help to find error

Mr. jeanpierreaubry
You are completely missing the point. I repeat.. the issue is - why one analysis runs and another do not run with almost same setting (yes.. the convergence settings are also same in both files).

Since same are the settings, either both files should run, or both files should show error. But only one is working while other is showing error. Hence, the confusion.

Last edited by nirmaljoshi (2019-05-22 07:11:07)

Offline

#8 2019-05-22 07:19:40

jeanpierreaubry
Guru
From: nantes (france)
Registered: 2009-03-12
Posts: 3,638

Re: [solved]help to find error

Anyways, can you probe the error in the "not-working" folder ? You may have already noted identical command in both files.

and i noticed some non identical things
be a bit more modest my friend
and read the .mess file

mesh in working

------------ MAILLAGE mesh     - IMPRESSIONS NIVEAU  1 ------------

ME-22-MAI -2019 05:42:19                                                        

NOMBRE DE NOEUDS                         570

NOMBRE DE MAILLES                       1480
                              SEG2                  432
                              QUAD4                 688
                              HEXA8                 360

NOMBRE DE GROUPES DE NOEUDS                3
                              Group_loadnodes                      5
                              Group_supportnodes                   5
                              Group_steelfibre                     1

NOMBRE DE GROUPES DE MAILLES               6
                              Group_load                           4
                              Group_support                        4
                              Group_concrete                     360
                              Group_botSteel                      36
                              Group_topSteel                      36
                              Group_symFace                       20

mesh in notworking

------------ MAILLAGE mesh     - IMPRESSIONS NIVEAU  1 ------------

ME-22-MAI -2019 06:01:15                                                        

NOMBRE DE NOEUDS                         660

NOMBRE DE MAILLES                       1804
                              SEG2                  588
                              QUAD4                 796
                              HEXA8                 420

NOMBRE DE GROUPES DE NOEUDS                5
                              Group_loadnodes                      5
                              Group_supportnodes                   5
                              Group_barendnodes                    4
                              GR_1_Group_steelfibre                1
                              Group_steelfibre                     1

NOMBRE DE GROUPES DE MAILLES               7
                              Group_concrete                     180
                              Group_topSteel                      42
                              Group_load                           4
                              Group_botSteel                      42
                              Group_support                        4
                              Group_symFace                       20
                              Group_shearSteel                    40

Last edited by jeanpierreaubry (2019-05-22 07:29:05)


Attachments:
notworkingmesh.png, Size: 23.48 KiB, Downloads: 44

consider reading my book
freely available here https://framabook.org/beginning-with-code_aster/

Offline

#9 2019-05-22 07:38:35

manonB
Member
From: Allemagne
Registered: 2019-04-02
Posts: 68

Re: [solved]help to find error

So I don't know if this can generate an eigenvalue problem, but in the .mess file you also have:

"Sur certaines mailles, la modélisation est incompatible avec le comportement.
Une erreur fatale pourrait suivre ce message."

Meaning that on some elements, the modelling (so in MODELISATION I guess) is not compatible with the behavior.

And I don't know if it is normal or not but you actually have 420 meshes with HEXA8 elements and you are affecting only 180 with 3D MODELISATION.

Offline

#10 2019-05-22 08:30:47

jeanpierreaubry
Guru
From: nantes (france)
Registered: 2009-03-12
Posts: 3,638

Re: [solved]help to find error

manonB wrote:

And I don't know if it is normal or not but you actually have 420 meshes with HEXA8 elements and you are affecting only 180 with 3D MODELISATION.

that is the point
the 2 mesh are quite different
as i advise in my book it is a good idea to use this line in the .comm file to view the mesh actually used in the calculation

IMPR_RESU(FORMAT='MED', UNITE=71,  RESU=_F(MAILLAGE=mesh,),);

consider reading my book
freely available here https://framabook.org/beginning-with-code_aster/

Offline

#11 2019-05-22 08:45:49

mecour
Member
From: Ostrava (Czech)
Registered: 2011-04-04
Posts: 153

Re: [solved]help to find error

Hello,

As Jean Pierre and manonB wrote you have badly created group in second case. The group is just created from one side surface Hexa elems. Another elements dont have assigned material.

Next time maybe try firstly check mesh


Attachments:
bad_group.png, Size: 12.69 KiB, Downloads: 71

Offline

#12 2019-05-22 12:47:16

nirmaljoshi
Member
From: Japan
Registered: 2018-10-12
Posts: 181

Re: [solved]help to find error

Thanks to all. I was stupid that I did not check the mesh properly before submitting here in the forum. Now everything is working fine.
Now, I have also modified the convergence criteria to required accuracy and the results are promising(i.e. more or less the result is matching with the experiment values). And of-course I have to tweak the parameter in the next step.

The comparison between 1d BARRE and 3d to model the reinforcement steel in a beam was the main objective of this analysis. Attached is the result.

Last edited by nirmaljoshi (2019-05-22 12:51:19)


Attachments:
testresults.png, Size: 13.49 KiB, Downloads: 105

Offline

#13 2019-05-22 16:41:40

jeanpierreaubry
Guru
From: nantes (france)
Registered: 2009-03-12
Posts: 3,638

Re: [solved]help to find error

when i have to solve several problems as you have here
i usually build only one mesh with all the groups
and i assign only the used ones in the various .comm files
or comment them at will


consider reading my book
freely available here https://framabook.org/beginning-with-code_aster/

Offline