Help needed: Calculation Error, "SPACER"

Hello again,

I tried to apply the 3-2-1 rule (one node fixed at three translations, one node at two, and one at one, + 1 surface loaded with the "pres" operator), without success. Same result ...
I also tried again to clamp one face + "pres" (as in the "PISTON" tutorial which works) , no success, same result. BTW: This were the boundaries I initially started from.

Then I did all the same on a different machine (openSUSE 10.3 with SALOME-MECA 2008.1 stand alone version), no success, same result.

My problem is that I do not even have a strategy to localize the problem, so there are a few questions...

- Are there critical geometry features of the part that could give such results?
- The mesh I use is rather coarse (-> for a first rough estimate). Does it make sense to try a (much) finer mesh ?
- Is there a difference between the available mesh algorithms concerning the stability of the calculation ?

All ideas welcome ...

Tom

Hi,

after days of search for the error I finally found it (at least I think so ... ):

The step file I used as starting point for the simulation resulted in a "compound" when imported into SALOME geometry module. Directly meshing the "compound" led to the described failure.

"Exploding" the compound into solids gave two single solids. These two solids I fused (boolean operation) and then I did the meshing, starting with the new, fused solid.

After setting the boundaries (1 point fixed in all 3 translational DOFs, and the support faces fixed in z-Direction with dz=0 + pressure applied on the designated face) the calculation case was carried out without any problems.

Thanks for your input and help,

best regards, Tom