CMS analysis in ANSYS Workbench 14.0

[MartinDU]

Hi,

I have a question regarding CMS analysis in ANSYS Workbench 14.0. I really hope that one of you experts out there can/is willing to help me.

I've tried normal substructuring (Guyan) and obtained the expected results in a modal analysis. Now, since the Guyan reduction yields significant deviations for most eigenmodes, I would like to try CMS. I've chosen a simple system, namely a 3-element cantilever beam:

|------o------o------o

I want to find the eigenmodes for this full model system by condensating the midmost element to a superelement, i.e. obtaining:

|------o SE o------o

The two midnodes are defined as remote points.
I have done the procedure in analogy to the Guyan reduction, i.e. first I have generated the superelement. This is done in a separate analysis on the basis of only the midmost element (the one transformed to superelement):

o------o

The two nodes are defined as remote points in the model. The generation command is (partially copied from a tutorial):

/FILNAME, BEAM2
/SOLU
/COM
NSEL, S, NODE, 1001 , N1
NSEL, A, NODE, 1002 , N2
M, ALL, ALL
ALLSEL, ALL
CMSEL, ALL
NUM_MODES = 100
/SOLU
ANTYPE, SUBSTR
SEOPT, BEAM2, 2
CMSOPT, FIX, NUM_MODES, 0, 100000
ESLN, S
SOLVE
FINISH
SAVE

Strangely, this doesn't result in an error as it did when using Guyan reduction (another command of course). But it generates a .sub file I've tried to use in a analysis of the reduced model (the one with the superelement in the middle). The usage command is:

/PREP7
ET,500,MATRIX50
TYPE,500
MAT,1
SE,BEAM2
FINISH
/SOLU

This command is taken from my Guyan reduction procedure. The analysis goes through, but I obtain completely wrong eigenfrequencies.
Can anyone tell me where my error(s) is/are? It would be highly appreciated.

Kind regards,
Martin Ulriksen
Student in structural engineering, Aalborg University, Esbjerg, Denmark.