-
In the optimization task editor, click the
Advanced tab.
-
Specify the Filter Radius of the Sigmund filter, which
smoothes the resulting optimization solution. Changing this value might help you
to avoid known problems from fluctuations in sensitivity values. The following
options allow you to specify the filter radius for a general bead optimization:
- Relative to average edge length
-
Enter the relative filter radius. Abaqus/CAE computes the filter radius as the value specified multiplied
by the average element length in the design area (the region
that will be optimized). The default value is four. A value of
zero turns off the filter; the resulting bead optimization might
generate an unsmooth result that is numerically optimal but not
a realistic physical solution.
- Absolute value
-
Enter the absolute value of the radius.
-
For the sensitivity calculation, choose whether the optimization
should consider only the nodes in the region that will be optimized (default)
or all the nodes in the model.
-
Enter the Bead perturbation.
Abaqus/CAE
uses this value to compute a semianalytical sensitivity analysis using a finite
difference on the element matrices. The finite difference is computed as the
perturbation value specified multiplied by the average element edge length. The
default value is 0.0001, which is suitable for most bead optimization problems.
-
Select Use
Abaqus
sensitivities when possible to use
Abaqus
to compute the design responses and their sensitivities whenever possible. This
workflow modification improves the optimization process performance.
-
You can choose Use Group Operator when possible to use
large groups of more than 5000 elements or nodes in the design response
definition in an efficient way. This workflow modification uses a new algorithm
based on Abaqus sensitivities.
|