Failed on enumerating of Faces

[UR_]

Hello,
there seems to be an issue, if more than one FemPressureConstrain object is used.

This is a box made from shells. There is one pressure object with two loads inside.

1 FemConstrainPressureObj with 2 loads.png (17.66 KiB) Viewed 1142 times

Result doesn't look so bad, despite faulty load/deflection directions (but that's another movie)

box of shells - 1 FemConstrainPressureObj with 2 loads.png (264.96 KiB) Viewed 1142 times

Now we split pressure loads into 2 objects.

2 FemConstrainPressureObjs 1 load each.png (20.17 KiB) Viewed 1142 times

Deflection is zero

box of shells - 2 FemConstrainPressureObj with 1 load each.png (102.5 KiB) Viewed 1142 times

In latter case, let's have a look into ccx's input file.
Both emitted node sets are identical. Therefore loads (+100 and -100) are zeroed out.
That should be avoided.

Used files:

boxunderpressure.zip

(189.93 KiB) Downloaded 20 times

Done with:
OS: Windows 10
Word size of OS: 64-bit
Word size of FreeCAD: 64-bit
Version: 0.17.10802 (Git)
Build type: Release
Branch: master
Hash: 10ce910c03347a90712327a7cc7be2bcdee13afc
Python version: 2.7.8
Qt version: 4.8.7
Coin version: 4.0.0a
OCC version: 7.0.0

[bernd]

Does it work for you with solid mesh?

[UR_]

Sorry, but solids are broken too

[bernd]

damn ... I'll have look ... We really need more unit tests. For sure this has been working with solid meshes ...

[bernd]

Sorry, but solids are broken too

In the regard of solid mesh. It works for me if

either:
- Netgen mesh object is used
or
- the GMSH mesh object is moved outside the analysis, the shape is meshed and the GMSH mesh object is moved back inside the analysis. (check the mesh object, it should not have any groups !)

What's the culprit ? The GMSH mesh object if used inside an analysis creates mesh groups at meshing. These groups are used for writing the input file. There is some bug. If the mesh object is meshed outside the analysis no groups are created and some algorithm is used to get the appropriate data for writing the input file. (BTW this algorithm is pretty fast thanks to ulrich ) There is no bug in this algorithm

[bernd]

this https://github.com/berndhahnebach/FreeC ... diff=split should fix it for all kind of pressure constraints with all kind of solid mesh objects.