Substructuring and modal analysis

[leoack00]

Hello everyone,
I need to perform a modal analysis of a large model with many stiffened panels but I'm not interested in the local modes hence I need to apply a guyan reduction. As I understand this is done using master nodes (M command) and ANTYPE, SUBSTR followed by ANTYPE, MODAL. I have read the advanced ansys.pdf and tried to follow the code but I keep getting the local modes. I have followed the top-down approach. The pdf mentions that I have to select all superelements, but as I know from FEMAP or NASTRAN, I should only have master nodes and not elements, mz master nodes are distributed all over the model. I understand this may also cause some problems later as I need to define the MATRIX50 elements but only have master nodes. Does this only work when dividing a model int separate but connected parts? Alternatively, is there a way to import a NASTRAN .dat file into APDL using PyMAPDL? I have the entire correct code with the substructuring as a .dat from NATSRAN. This is the code I have used so far:

/FILNAME,FULL
...... (building entire model)
CM, master_N, NODE         !Select all master nodes to create CM with master nodes

NSEL, ALL
EPLOT

SAVE
FINISH

/FILNAME,GEN

/PREP7
ANTYPE,SUBSTR
SEOPT,GEN,3,0,0,RESOLVE
FINISH

/SOLU
CMSEL,S,master_N,NODE
M,master_N, UX, , UY, UZ
SOLVE
FINISH

EPLOT

/CLEAR
/FILNAME,USE
RESUME,FULL,DB

/PREP7
ET,99999,MATRIX50
TYPE,99999
SE,GEN
EPLOT
FINISH

/SOLU
ANTYPE,MODAL
MODOPT,LANB,20
SOLVE
FINISH

/CLEAR
/FILNAME,GEN
RESUME,FULL,DB
/PREP7
ANTYPE,SUBSTR
SEEXP,GEN,USE
EXPASS,ON
!EXPSOL,ALL			! Is this useful?

/SOLU
SOLVE
FINISH

[mikerife]

Hi @leoack00
Please open an traditional Ansys Inc technical support case since you are really asking about a Mechanical APDL workflow. I do see two problems right away; the M command is not correct as there is a "UY" in the NINC field: NINC is the increment of nodes numbers. Also we don't resume the full model database on the Use pass step. Well at least not without steps that would 'get rid of' the FEM in the full model that is represented by the substructure. In this case it looks like the entire 'full' model was turned into a substructure. Unless you just skipped copy/pasting a lot of commands.

Mike