Analyzing the Optimization Results

In this optimization design run, the SQP algorithm searches for the best parameter combinations, in the specified range, that result in a local minimum of the pressure drop.

To analyze the outcome of the optimization study, you can create a history plot of the pressure drop. When the pressure drop reaches a local minimum, the sensitivities of the parameters also approach their minimum. Low sensitivity indicates that changing the value of the parameter around its current value does not reduce the pressure drop rapidly. The lowest sensitivity is an indicator that the local minimum is reached.
  1. To create a history plot for the pressure drop:
    1. Right-click Design Studies > Design Study node and select Create Plot > History.
    2. In the History Plot Setup dialog, use the default properties as follows:
      PropertySetting
      Data SourceDesign Study
      Design SetsAll
      Left Axis DataResponses > dP
    3. To adjust the plot range, select the History Plot > Axes > Left Axis node and set Maximum to 200.0.
    The optimization history plot appears as follows:


    After 10 SQP outer iterations, the pressure drop reduces from 123.2 Pa to 59.7 Pa. If you wish to continue the SQP search, you can increase the Maximum Outer Iterations to 15.

  2. To create a sensitivity plot for all the design parameters:
    1. Right-click the Design Studies > Design Study node and select Create Plot > Sensitivity.
    2. In the Sensitivity Plot Setup dialog, use the default properties as follows:
      Property Setting
      Design Study Design Study
      Design Sets All
      Left Axis Data

      dP

      P1-v1-y,P1-v2-x, P1-v3-x,… (select all the geometric parameters)


    If the pressure drop approaches to the local minimum, the sensitivities of the geometric parameters converge to stable small values.

    The optimization stops when the Maximum Outer Iterations value reaches the specified value of 10.