$ mmfindfix1 example of massless mechanisms $ see notes after enddata nastran system(359)=0 $ block switch from lan to ahou for small problem DIAG 8 SOL 103 $ Include 'mmfindfixa.v2001' $ CEND TITLE = massless mechanism example mmfindfix1 Subtitle = test all paths through rb discard logic ECHO = BOTH METHOD = 1 Disp = all Subcase 1 Label = no rb modes Spc = 1 Bc = 1 Subcase 2 Label = one massless mechanism Spc = 2 Bc = 2 Subcase 3 Label = one MM, one rb mode Spc = 3 Bc = 3 Subcase 4 Label = two MM, one rb mode Spc = 4 Bc = 4 Subcase 5 Label = 2 rb modes Spc = 5 Bc = 5 BEGIN BULK $Param, usetprt, 0 SPC1, 1, 123456 11 101 $ no rb mode SPC1, 2, 12356 11 101 $ torsion MM SPC1, 3, 2356 11 Spc1, 3, 1, 101 $ axial rb mode, torsion MM SPC1, 4, 2356 11 $ two MM, torsion and disjoint mode, $ axial rb mode SPC1, 5, 3456 11 $ axial, lateral rb modes Spc1, 5, 1, 101 $ axial rb mode, torsion MM Param, coupmass, 0 Param, mechfix, debug $ force mechfix to run, with more diagnostics Param, mechprt, yes $ print constraint modes as spcforces GRID, 11, , 1. GRID, 12, , 2. GRID, 13, , 3. CBAR, 112, 112, 11 12 0. 0. 1. CBAR, 123, 112, 12 13 0. 0. 1. $Conm2, 13, 13, , 0.0 $ define a torsion rb mode $, 1.0 $ I11 PBAR, 112, 11, 1. 2. 3. 4. MAT1, 11, 1.+10, , .3, 1. $ use large value of E to separate rb mode freqs from flex mode freqs. EIGRL, 1, , , 6 $ GET 6 MODES $ disjoint structure, to move size past autoswitch. Very stiff, light, no $ effect on first 4 modes. Used to avoid autoswitch to AHOU method. $SPOINT, 1001 THRU 1020 $celas2, 1001, 1.e6 1001 $=, *1, =, *1 $=16 $cmass2, 2001, 1.e-2 1001 $=, *1, =, *1 $=16 $ add a second massless mechanism grid, 101 grid, 102 celas2, 10102, 1., 101, 1, 102, 1 ENDDATA Error report Massless Mechanism Constraint-Constraint of rigid body modes. The massless mechanism feature (MM) introduced in V2001 first finds DOFs with high maximum ratios in matrix decomposition for [K + s*M]. It then finds shapes associated with these high ratio DOFs. These shapes may also include rigid body modes. This is particularly true when the first flexible mode is much greater than 10.0 Hz. The feature removes the rigid body modes, then constrains only the true massless mechanisms. There is a logic error in v2001 that does not discard shapes that are rigid body modes, when there are no massless mechanisms. The alter package mmfindfix.v2001 fixes the error. It is available on MSC's web site in the SSSALTER directory.