Setting Calibration Options

You can specify error measures, optimization and execution controls, and request parameter sensitivity calculations for each calibration.

See Also
Best-Fit Error Measures
Calibration Algorithms and Parameter Sensitivity
  1. From the Calibration Tools section of the action bar, click Calibration Setup .
    The Calibration Setup dialog box appears.
  2. From the commands section at the top of the Calibration Setup dialog box, click Optimization Controls
    The Optimization Controls dialog box appears.
  3. Specify the error measure behavior from the Error measure settings options.
    1. Specify a Best-Fit Error Measure algorithm for measuring the deviation between test data and the projected material behavior.
    2. Select Normalize error to specify the normalized version of the error measure algorithm.
  4. Specify the basic optimization behavior for the calibration.
    1. Expand the Basic optimization settings options.
    2. Specify an Optimization algorithm and, if applicable, adjust the other optimization settings. You can adjust the following settings that apply to all iterative solutions:

      Setting Description
      Solution tolerance Controls the convergence criterion based on the change in calibrated parameters during consecutive iterations.
      Maximum function evalutions Specifies the maximum allowed number of function evaluations.
      Function tolerance Controls the convergence criterion based on the decrease in the objective function value during consecutive iterations
      Maximum iterations Specifies the maximum allowed number of iterations.

    3. Select Compute parameter sensitivities to compute parameter sensitivities during the calibration.
    4. If the numerical execution kernel is active, set the Number of CPUs you want to use for the calibration.

      Using multiple CPUs during a calibration can reduce the overall wall clock time. The number of CPUs you use should be less than or equal to the number of test data sets to be used during calibration.

  5. Optional: Set advanced optimization options.
    1. Expand the Advanced optimization settings.
    2. Modify the values for any of the algorithm specific advanced optimization settings that drive optimization behavior.
    3. (Particle swarm algorithm only) Select Use local topology to propagate the information about the objective function decreasing from a particle to the particles in its neighborhood only. Leave this option clear to communicate the information about the best achieved location to all particles in the swarm.
  6. Optional: Modify default tolerances and solver settings:
    1. If the numerical execution kernel is active, expand the Advanced solver settings options, and specify settings required for your calibration.
    2. Modify the values for any of the solver behavior settings, including maximum strain increment and relative tolerance.
  7. Optional: Expand the Units options, and specify the units that you want to use for the calibration.
  8. Close the panel.