Analyzing optimization results

After the optimization solve is finished, you can view results from each iteration in the Post Processing Navigator or in the optimization spreadsheet.

Post Processing Navigator

By default, OP2 results for each iteration are saved for analysis. You can display contour plots from each iteration using the Post Processing Navigator.

When you animate optimization results using the Animation command, set the Animate option to Iterations to visualize the changing mesh in each iteration. You can specify the range of optimization design cycles that the animation includes.

Optimization spreadsheet

After the optimization solve is finished, design variable values for each iteration are saved in a spreadsheet. To open the spreadsheet:

• In the Simulation Navigator, right-click the optimization solution process and choose Optimization Spreadsheet or Sensitivity Spreadsheet.The Sensitivity Spreadsheet command is available only if you solved a global sensitivity study.

The spreadsheet lists all attempted iterations and displays failed design constraints results in red.

Spreadsheet results available for optimization include:

• Design Objective Function Results—The name of the function is displayed along with its value for each iteration.

• Design Variable Results—The name of each design variable is displayed along with a result value for each iteration.

• Design Constraint Results—The name of each constraint variable is displayed along with a result value for each iteration.

Note:

The best design is not necessarily the last design iteration, even if the optimization converged.

A design iteration is considered feasible if all constraint values in the iteration remained within their limits while the objective function reaches toward its intended goal (such as minimizing or maximizing a target value). Constraint values that violate their limits appear in red in the spreadsheet.

If there are one or more feasible design iterations (as indicated by one or more iteration columns with all black values for the design constraint results), then the best design is the iteration whose design objective function result most closely meets its intended goal.

If there are no feasible designs (that is, at least one red value appears in each column in the design constraints results section), then the best infeasible design is the iteration with the least violation of the most critical constraint, without regard to the value of the objective function.

If the results are not satisfactory, you can tighten the convergence parameter values (which will require more iterations to converge) or set the model to the most feasible design iteration values as a new starting point for the optimization analysis.

How do I

Create the geometry optimization

Create a weight or volume objective

Create a frequency objective

Create an objective from a result measure

Create a weight or volume constraint

Define a frequency design constraint

Create a constraint based on a result measure

Define design variables for the geometry optimization

Solve the geometry optimization

Learn more

Geometry optimization

Geometry Optimization workflow

Types of geometry optimization

Design objective – Geometry Optimization

Defining design constraints – Geometry Optimization

Defining design variables – Geometry Optimization

Part files modified by Geometry Optimization

Update an expression with the result from a specific iteration

Look up more details

Geometry optimization parameters

Quick links

Exporting result quantities as named expressions

Command reference

Pre/Post video examples

Bulk Entry Descriptions

Simcenter 3D tutorials

Browse Simcenter 3D help by product area

Analyzing optimization results, Simcenter 3D 2021.1 Series

© 2020 Siemens

window.mainLanguage="en_US"

window.delivId=""

window.projectId=""

MathJax.Hub.Config({ TeX: { extensions: ["autoload-all.js"] }, tex2jax: { displayMath: [ ] }, "SVG": { scale: 125 } });