Generating Frequency-Based Substructures

This section describes how to define individual frequency-based substructures.

Frequency-based substructures differ from conventional substructures (see Generating substructures) in that they use stiffness, inertial, damping, and frequency information to generate a single complex, condensed operator at every specified frequency point.

You can use frequency-based substructures only in direct steady-state dynamic analyses (see Direct-Solution Steady-State Dynamic Analysis).

For information on how substructures are used in a model, see Using Substructures.

This page discusses:

See Also
Using Substructures
In Other Guides
*RETAINED NODAL DOFS
*STEADY STATE DYNAMICS
*SELECT EIGENMODES
*RESTART

Products Abaqus/Standard

You define a frequency-based substructure using a combination of a direct steady-state dynamic analysis and a substructure generation procedure. The direct steady-state dynamic analysis coupled with retained external degrees of freedom generates the operators corresponding to the frequency-based substructure at the user-specified frequencies. The substructure generation procedure stores both the frequency-based substructure operators and the conventional substructure operators on the merged substructure database. One or the other set of operators is used depending on the analysis type and substructure property settings.

All general behaviors related to substructure generation and usage apply to frequency-based substructures. This section discusses details that are pertinent to frequency-based substructure generation.

Frequency-based substructures work in direct steady-state dynamic analyses in the same way that conventional substructures work in static analyses. Conventional substructures provide exact (up to round-off error) representation of the model in static analyses, while frequency-based substructures provide exact representation of the model in direct steady-state dynamic analyses. When the frequency-based substructure is used at the same frequencies as those used for generation, the dynamic response is exact (up to round-off error). The response at frequencies that do not match those used for generation can be approximate (see Using Substructures).

To define frequency-based substructures, do the following:

  1. Invoke the direct steady-state dynamic procedure.

    1. Define the frequency points at which to generate a frequency-based substructure.

    2. Define the nodes and degrees of freedom to retain as external degrees of freedom when using the frequency-based substructure.

  2. Invoke the substructure generation procedure.

    1. Define the nodes and degrees of freedom to retain as external degrees of freedom when using the frequency-based substructure. These degrees of freedom must be the same as those specified in the direct steady-state dynamic procedure.

    2. Define the rest of the substructure generation procedure (see Generating substructures).

You can define frequency-based substructures using a single analysis or using the restart capability. Because the substructure generation procedure runs only in shared memory parallel (SMP) mode, the performance can be limited when the direct steady-state dynamic analysis and substructure generation procedure are run in a single analysis. You can use the restart capability to run two separate jobs and allow the direct steady-state dynamic analysis to run in distributed memory parallel (DMP) mode to reduce the computational expense, particularly for large models.