[Getdp] magnetostatic 3D, how to improve the repeatability ?

[Joseph Frippiat]

Hello,

I am playing with a small magnetostatic 3D problem (the files are 
attached to this message) and I never get two times the same results 
when I run it.  Even if I refine the meshing.

If the meshing is too small (when I set Mesh.CharacteristicLengthFactor 
to 1.0 in the file OneTooth.geo), I get this error message:

GetDP   : Selected Resolution 'MagSta_a_3D'
GetDP   : Loading Geometric data 
'M:\JFr-mobile\Getdp\studies\griffes\VerySimpleBrake\magnetostatic\OneTooth.msh'
GetDP   : System 'Sys_Mag' : Real
GetDP   : P r e - P r o c e s s i n g . . .
GetDP   : Treatment Formulation 'Magnetostatics_a_3D'
GetDP   :   Generate ExtendedGroup '_CO_Entity_11' (EdgesOf)
GetDP   :   Generate ExtendedGroup '_CO_Entity_12' (EdgesOf)
GetDP   : (CPU = 2.23081s)
GetDP   : E n d   P r e - P r o c e s s i n g
GetDP   : P r o c e s s i n g . . .
GetDP   : Generate[Sys_Mag]
GetDP   : (CPU = 3.82202s)
GetDP   : Solve[Sys_Mag]
GetDP   : N: 162201 - preonly lu mumps
Error   : Did not receive message body: stopping server
Info    : Done running 'GetDP'

What do I need to change or correct to improve the simulation ?

Thanks for any help,

Joseph

-------------- next part --------------
A non-text attachment was scrubbed...
Name: OneTooth.zip
Type: application/x-zip-compressed
Size: 6114 bytes
Desc: not available
URL: <http://www.geuz.org/pipermail/getdp/attachments/20120818/a20e6300/attachment.zip>