×

Notice

The forum is in read only mode.

General CAELinux/Salome Questions

  • Brian Hale
  • Topic Author
  • Offline
  • New Member
  • New Member
More
11 years 1 month ago #7445 by Brian Hale
General CAELinux/Salome Questions was created by Brian Hale
Hey all,

First. to all the people who helped make this software, THANK YOU!, I bow before your prowess,

I am new to the Linux/CAELinux and so far I am really liking it but I have some general questions. (if I need to break these into separate posts just let me know.)

1. Most big brand FEA software have a way to check for convergence of a solution, is there a way to do this in any CAELinux applications (eg. Salome)?

2. Has anyone done any kind of electronics cooling analysis using CAELinux, if so any suggestions on where to find help/tutorials?

3. In general FEA I am familiar with the term to add "mid-side nodes" and what adding them does, but in Salome what does it mean to "convert to quadratic" or are these synonymous.

4. In other big brand software there is the ability to generate a push button report that will list constraints, loads, mesh info, results, etc.. Does any CAELinux software have this ability?

And just so I am clear, non of these items are deal breakers for the software. If CAELinux doesn't do any of these it is still great software that I will continue to use, but as I am new to Linux, CAELinux, and french :cheer: I just wanted to make sure I am not missing anything.

Thanks for your time
More
11 years 1 month ago #7448 by Claus
Replied by Claus on topic Re: General CAELinux/Salome Questions
Okay, I'll give it a shot and answer some of your questions:

1. Not as far as I know, but that does certainly not mean there aren't any. When you get to know the code, you will start to get a feel for for the validity of your solution and you will instinctively know when you wont gain anything more from mesh refinement.
What convergence criteria are you thinking of?

2. There are several thermal analysis packages in CAELinux, so take your pick :)

3. Yes, as far as I know, you can call them synonymous. There are variations of quadratic elements via Salome and/or via Code_Aster command file.

4. For Salome/Code_Aster unfortunately no. I not that familiar with the other software included though.

/C

Code_Aster release : STA11.4 on OpenSUSE 12.3 64 bits - EDF/Intel version
  • Brian Hale
  • Topic Author
  • Offline
  • New Member
  • New Member
More
11 years 1 month ago #7449 by Brian Hale
Replied by Brian Hale on topic Re: General CAELinux/Salome Questions
WOW a reply from the great and powerful Claws! :woohoo: your tutorials and posts have been quite helpful to me thanks.

1. In Comsol and Ansys (for example) there is a plot that is generated when when you run your analysis and as long as the slope of that curve is negative (downward) toward a singular value (ie. converges in stead of diverges) then you are reasonably sure you have a good answer. Other than that I am really not too sure how it works or what it is doing. Currently I just keep decreasing the size of my mesh (increasing the number of nodes) and see if my solutions change, if they are within about 3% or less of each other I consider it to have converged. Right or wrong it is about the only method I have and it is pretty time consuming.

2. Can they all do CFD Heat Transfer? I thought most of them couldn't.

3. So in Salome mesh module, when I right click>convert to/from quadratic and do not check the create middle node flag and click convert, what is it doing?

4. Thats what I thought thanks.
More
11 years 1 month ago - 11 years 1 month ago #7451 by kwou
Replied by kwou on topic Re: General CAELinux/Salome Questions
Hoi Razzack

3. So in Salome mesh module, when I right click>convert to/from quadratic and do not check the create middle node flag and click convert, what is it doing?


when ''convert to/from quadratic'' with ''medium nodes on geometry''
- checked: medium nodes follow the geometry (so eg. elements on a circle remain on the circle)
- unchecked: medium nodes are on a straight line between the neighbouring nodes.
It seems I cannot upload files here, so have a look there:
www.caelinux.org/wiki/index.php/Contrib:KeesWouters/Tests


Convergence check:
In the output window and later in the *.mess file the residues (both relative and absolute) are given during/of the iteration. See the non-converged solution below for an analysis with hyper elastic material (Mooney-Rivlin):
Instant de calcul:  1.020123456790e-01 
 
---------------------------------------------------------------------
|   ITERATIONS   |     RESIDU     |     RESIDU     |     OPTION     |
|     NEWTON     |     RELATIF    |     ABSOLU     |   ASSEMBLAGE   |
|                | RESI_GLOB_RELA | RESI_GLOB_MAXI |                |
---------------------------------------------------------------------
|     0        X | 1.81926E-05  X | 1.53798E-06    |TANGENTE        |
|     1        X | 3.44886E-06  X | 2.91564E-07    |TANGENTE        |
|     2        X | 3.12011E-05  X | 2.63774E-06    |TANGENTE        |
|     3        X | 1.93978E-05  X | 1.63994E-06    |TANGENTE        |
|     4        X | 1.81436E-04  X | 1.53401E-05    |TANGENTE        |
|     5        X | 1.11695E-04  X | 9.44510E-06    |TANGENTE        |
|     6        X | 1.11275E-03  X | 9.41308E-05    |TANGENTE        |
|     7        X | 6.46989E-04  X | 5.47821E-05    |TANGENTE        |
|     8        X | 1.81633E-02  X | 1.54121E-03    |TANGENTE        |
|     9        X | 3.48166E-03  X | 2.96912E-04    |TANGENTE        |
|    10        X | 5.43568E-01  X | 4.69019E-02    |TANGENTE        |
|    11        X | 1.75774E-02  X | 1.54306E-03    |TANGENTE        |
|    12        X | 1.14018E+01  X | 1.06691E+00    |TANGENTE        |
|    13        X | 1.25460E-01  X | 1.24214E-02    |TANGENTE        |
|    14        X | 5.99510E+01  X | 5.86890E+00    |TANGENTE        |
|    15        X | 2.15687E-01  X | 2.32612E-02    |TANGENTE        |
|    16        X | 4.41877E+01  X | 4.00688E+00    |TANGENTE        |
|    17        X | 4.57658E-01  X | 4.07324E-02    |TANGENTE        |
|    18        X | 2.13911E+00  X | 2.05772E-01    |TANGENTE        |
|    19        X | 3.02427E-01  X | 2.95475E-02    |TANGENTE        |
---------------------------------------------------------------------                                                                                                                                                

 <Erreur> Le nombre maximum d'itérations est atteint
 
|    20        X | 4.37819E+01  X | 4.25189E+00    |TANGENTE        |

  Subdivision du pas de temps en 5 sous-pas
Méthode de subdivision : PENDANT LA NOUVELLE SUBDIVISION
Le pas minimal de la subdivision est atteint.
   Pas de Temps actuel          = <1.371742E-06>
   Pas de Temps minimum imposé  = <1.000000E-05>
   Niveau de subdivision        = <0>
   Méthode de subdivision       = <UNIFORME>
Conseil :
   Diminuer SUBD_PAS_MINI. Il est également possible d'ajuster SUBD_NIVEAU pour
   indiquer le nombre successif de subdivision d'un pas de temps.
   Si les 2 mots clefs SUBD_NIVEAU et SUBD_PAS_MINI sont utilisés la subdivision
   s'arrête dès que l'un des 2 critères est vérifié.

Interest: structural mechanics, solar energy (picture at 'my location' shows too little pv panels)

--
kind regards - kees
Last edit: 11 years 1 month ago by kwou.
  • Brian Hale
  • Topic Author
  • Offline
  • New Member
  • New Member
More
11 years 1 month ago #7452 by Brian Hale
Replied by Brian Hale on topic Re: General CAELinux/Salome Questions
Keeswouters,

You have turned on two light bulbs but turned off one :laugh: (bad joke I know)

Your quadratic answer makes perfect sense. But then why do I get a warning stating "some nodes are not placed on geometry...." when I check "medium nodes on geometry"? I might have to go back and take a look at this.
I assumed it had something to do with the <CALCULEL_7> error I kept getting......I think i am more confused by my analysis now but thank you for the clarity of your answer.

I love the convergence check, but is that information something you only get when you do a non-linear analysis? I went back and checked my linear-static.mess file and I do not have a table of numbers like that. I do have a table but it looks like this:
PARTITION :    1
--------------------------------------------------------------------------------

 CL-  --NUM-- -MA-  -----IADM----- -U- - LON UA -  -S- ------------- NOM --------------

|G|   0|       0|   0|           9|X|       7438| X| <<<<         LIBRE          >>>>
| |   0|       0|   0|        7455|U|        250| D|                                 
| |   0|       0|   0|        7713|U|         50| D|                                 
| |   0|       0|   0|        7771|X|     123294| X| <<<<         LIBRE          >>>>
--------------------------------------------------------------------------------
OBJETS ALLOUES DYNAMIQUEMENT 
--------------------------------------------------------------------------------

 CL-  --NUM-- -MA-  ---------IADY--------- -U- - LON UA -  -S- ------------- NOM --------------

|G|   0|       1|  -2|           176694256|U|         11| D| ________GLOBALE ________$$CARA  
|G|   0|       2|  -2|           144075872|U|       4000| D| ________GLOBALE ________$$IADD  
|G|   0|       3|  -2|           141819872|U|        251| D| ________GLOBALE ________$$GENR  
|G|   0|       4|  -2|           141821968|U|        251| D| ________GLOBALE ________$$TYPE  
|G|   0|       5|  -2|           141839952|U|       1001| D| ________GLOBALE ________$$DOCU  
|G|   0|       6|  -2|           144107952|U|       2000| D| ________GLOBALE ________$$ORIG  
|G|   0|       7|  -2|           144124032|U|       8004| D| ________GLOBALE ________$$RNOM  
|G|   0|       8|  -2|           144188144|U|       2000| D| ________GLOBALE ________$$LTYP  
|G|   0|       9|  -2|           144204224|U|       2000| D| ________GLOBALE ________$$LONG  
|G|   0|      10|  -2|           144220304|U|       2000| D| ________GLOBALE ________$$LONO
.
.
.
.
.

Or it could just be that my analysis is that messed up!
More
11 years 1 month ago - 11 years 1 month ago #7456 by kwou
Replied by kwou on topic Re: General CAELinux/Salome Questions
Hoi Razzack

You have turned on two light bulbs but turned off one (bad joke I know)

The question then is: how many bulbs are turned on after the joke?

Your quadratic answer makes perfect sense. But then why do I get a warning stating "some nodes are not placed on geometry...." when I check "medium nodes on geometry"? I might have to go back and take a look at this.
I assumed it had something to do with the <CALCULEL_7> error I kept getting......I think i am more confused by my analysis now but thank you for the clarity of your answer.

I am not sure I understand your remark completely:

... I get a warning stating "some nodes are not placed on geometry...." when I check "medium nodes on geometry"? .

This is reported by Salome?

the <CALCULEL_7> error I kept getting....

This is reported by Code Aster? So why should these errors/remarks be related?

I love the convergence check, but is that information something you only get when you do a non-linear analysis?

The short answer is: yes.

I like short answers, but I cannot refuse:
Ansys only throws the nice (interactive) convergence graph during a non-linear analsyis, if I am not mistaken?
In the out file there is no information on convergence either for a linear analysis. Correct?

Interest: structural mechanics, solar energy (picture at 'my location' shows too little pv panels)

--
kind regards - kees
Last edit: 11 years 1 month ago by kwou.
Moderators: catux
Time to create page: 0.143 seconds
Powered by Kunena Forum