A good test-analysis model (TAM) is created in an iterative process. The user first chooses a set of instrumented DOF, possibly using the techniques described in Section 2.5. Based upon the methodology presented in Chapter 2, a TAM is then created using the analysis techniques described in this section. The adequacy of the TAM is assessed using techniques which were described in Chapter 2.4. If the TAM is not satisfactory, the set of retained DOF is revised and a new TAM created until a satisfactory TAM is achieved.
The procedures needed to create a TAM depend on the method of reduction. The static or Guyan reduction requires only a single run with no special purpose alters, though alters may be required to evaluate the accuracy of the TAM. All other methods require the use of a DMAP alter and some of them also require the FEM mode shapes. The TAM procedures which are common to all TAM methods are described in this section. The method-specific requirements of the various TAM procedures along with example input files are described in Sections 3.2.1 through 3.2.6. For all these cases, the DOF set developed using the RKE method and presented in Table 3-13 is used.
The requirements which are common to all TAM methods are listed below for the various portions of an MSC/NASTRAN input deck.
Executive Control Deck
· SOL 103 must be selected.
Case Control Deck
· Boundary conditions (if any) must be selected by MPC and/or SPC cards.
· A METHOD card must be included to select the eigenvalue solution method. The eigensolution will be performed using the reduced TAM stiffness and mass matrices. The results of the eigensolution provide an extremely important check of the accuracy of the TAM reduction process.
· Generalized dynamic reduction (GDR) must not be used. Do not include a DYNRED card in the Case Control deck[13].
· Displacements may be recovered by including a DISP output request (such as DISP(PLOT)=ALL) in the Case Control deck. The mode shapes can be plotted using I‑DEAS or any other postprocessor to provide additional insight into any differences between the FEM and TAM results.
Bulk Data Deck
· The DOF to be retained in the TAM must be identified by ASET or ASET1 cards. For a superelement run, all A‑set DOF must be interior to the residual structure.
· Generalized dynamic reduction must not be used. Do not include SEQSET or QSET cards for upstream superelements or the residual structure, respectively.
· All standard bulk data (grids, elements, properties, etc.) which define the finite element model must be included.
· PARAM,USETPRT,0 may be used to monitor user sets.