Optimetrics
Use Optimetrics to determine the best design variation among a model's possible variations. You create the original model, the nominal design, then define the design parameters that vary, which can be nearly any design parameter assigned a numeric value in Twin Builder. For example, you can parameterize the model geometry or material properties. You can then perform the following types of analyses on your nominal Twin Builder design:
|
In a parametric analysis, you define one or more variable sweep definitions, each specifying a series of variable values within a range. For example, you can parameterize component values. (See Working with Variables for more information.) Optimetrics solves the design at each variation. You can then compare the results to determine how each design variation affects the performance of the design. Parametric analyses are often used as precursors to optimization solutions because they help to determine a reasonable range of variable values for the optimization analysis. |
|
|
For an optimization analysis, you identify the cost function and the optimization goal. Optimetrics changes the design parameter values to meet that goal. The cost function can be based on any solution quantity that Twin Builder can compute. |
|
|
In a sensitivity analysis, you use Optimetrics to explore the vicinity of the design point to determine the sensitivity of the design to small changes in variables. |
|
|
Tuning lets you change variable values interactively while monitoring the performance of the design. If you want to ensure that tuning does not resolve variations already solved by parametric setup, select Save Fields Mesh in the Options tab of the Optimetrics setup. |
|
|
In a statistical analysis, you use Optimetrics to determine the distribution of a design's performance, which is caused by a statistical distribution of variable values. |
|
| Design of Experiments | Design of experiments involves the creation of a response surface that depicts and predicts how variable values will affect the design. |
| DesignXplorer | An optimization tool for studying a range of design variations, used with design of experiments. |
Sweeping or using a complex variable is not allowed in any optimetrics setup, including optimization, statistical, sensitivity, and tuning setups.
Related Topics
Setting up a Parametric Analysis
Setting up an Optimization Analysis
Setting up a Sensitivity Analysis