Hi,
It’s not clear from the manual if *equation stated inbetween two degrees of freedom refers the global or local axis.
I’m particularly interested when used with beams and understand why rotational degrees of freedom sometimes do not couple properly with equation .
Regards
At least Abaqus can use local axes in equation constraints:
If a local coordinate system is defined for any node involved in the equation, the variables at that node appear in the equation in the local system.
But in CalculiX we can expect all kinds of quirks in such cases so it would require some testing.
This comes without any kind of warranty of what must be true or false, but only with reference to this small example with C3D6 element
*Node
1, 0.00, 00.0,-0.01
2, 0.05,-0.01,-0.01
3, 0.05, 0.01,-0.01
4, 0.00, 0.00, 0.01
5, 0.05,-0.01, 0.01
6, 0.05, 0.01, 0.01
*Element, Type = C3D6, Elset = Eall
1, 1,2,3,4,5,6
**
*Nset, Nset = Centernodes
1,4
*Nset, Nset = Base
1
*Nset, Nset = Cylnodes
2,3,5,6
**
*Material, Name=PET
*Density
1.42E-09
*Elastic
2960, 0.37
**
*Solid section, Elset=Eall, Material=PET
**
*Step, Nlgeom, Inc=100
*Static, Solver=Pardiso
0.1, 1, 1E-05, 0.1
*Transform, Type=C, Nset=Cylnodes
0, 0, 0, 0, 0, 1
*Boundary
Cylnodes, 2, 2, 0
Cylnodes, 1, 1, 0.01
*Boundary
Centernodes, 1, 2, 0
Base, 3, 3, 0
*Boundary, User
Cylnodes,1,3,1
**
*Node file
RF, U
*El file
S, E
*End Step
by using the uboun.f subroutine to list node numbers one will notice, that running without the *Transform card, then the cylnodes will appear as
input = internal → 2,3,5,6 = 2,3,5,6
but with the *Transform card then the cylnodes will be mapped into a new set of numbers
input = internal → 2,3,5,6 = 7,8,9,10
so still without warranty my best guess will be, that some of the equation node numbers aren’t mapped into the new node numbers assign by the transform card.
Thanks both
I will be doing some more tests this days to see if I can find what’s going on.