Looking for a help in translation (french-english)

Hello,
When trying to solve a static mechanical 2D plane model, the solver fails. And I can't work out what the error message means. Can someone translate it into english, please:

!_______________________________!
! !
! !
! !
! Matrice non factorisable: !
! pivot presque nul a la ligne: 13000 !
! nombre de decimales perdues: 13 !
! pour le noeud N6482 et la composante DY !
!
!
! Cette erreur est fetale. Le code s'arrete.!
!_______________________________!

Thanks in advance,
Vito<br><br>Post edited by: vito, at: 2007/09/12 11:53

it means (about) that the decompostion of the stiffness matrix fails and the Code stops due to the fatal error.
Johannes

Hi, you need to add boundary condition to solve your problem. The solver is unable to complete his work because there is a \&quot;zero\&quot; or \&quot;near zero\&quot; into the \&quot;diagonnal\&quot; of the matrix. Maybe you will have to add a boundary condition like \&quot;no translation\&quot; in Y direction at node 6482.

It can be another problem if you have crack element into your model. Sometime crack elements have a very big displacement and can be almost impossible to solve because of the difference between the rigidity of crack and the rest of the model.

I hope it will help you.

Yves

I've had this same message

&quot;&lt;F&gt; &lt;FACTOR_21&gt; !
! !
! !
! Matrice non factorisable : !
! pivot vraiment nul &amp;#65533;la ligne : 372376 !
! pour le noeud N2G92 et la composante DX&quot;

Which I translated via Babelfish (babelfish.altavista.com/tr)

&quot;! Stamp not factorisable: !
! pivot really no one the line: 372376!
! for node N2G92 and component DX!&quot;

But how to find that node referenced?... Salome does not recognize a node of name &quot;N2G92&quot; to locate it.

In the ASTK run screen, open the x.comm file in EFICAS and edit the line &quot;LIRE_MAILLAGE&quot; to add &quot;INFO&quot; with a value of &quot;2&quot;
(inferred from this link 64.233.169.104/search?q=cache:NakomrBRDT...k&cd=1&gl=us)

it's running now with a huge output file (I have a complex part). If I don't run out of disk space it looks like I can find the coordinates for the node &quot;N2G92&quot; and work backward to see what is going on and fix the problem (will probably have to involve a boundary condition for the X-axis at this node, but no way to find out which was referenced.

J

Update:... The file x.mess will now have the full list of nodes in this format: &quot;121154&quot;, &quot;N2G92&quot;, X Y Z coordinates. The numerical node number (first listed) can be put in Salome/Mesh/&quot;move node&quot; option and you'll see where it is (the coordinates should match from the file if you copied the right number), if it's hard to see then move it 10mm or 100mm to make it stand out to find it. If you're in the geometry module then create a small block and translate it to the coordinates of the error. My case I found a sliver of floating material as the problem (left over from geometry construction), but could have been bad boundary condition setup as others noted.<br /><br />Post edited by: J S, at: 2008/05/30 18:27