Using a Field as a Field-MOP Input to Generate RF Amplitudes in oSP3D

A field or signal quantity of interest may depend on other fields or signals. For the field of interest, you can create a field-MOP that takes a field as an input.

With a random field model of the input field, you can represent a particular realization of the input field with a small number of scalar random field amplitude parameters. You can then use these parameters directly as field-MOP inputs.


Note:  A field quantity with an existing random field is required.


Generating Amplitudes from Random Field Data

To generate amplitudes from random field data:

  1. Import field designs for a field quantity with an existing random field model.

  2. From the list of field quantities, select this quantity.

  3. Select Field data models > Approximate from imported data > Amplitudes from random field data.

oSP3D adds new scalar random field amplitude to the database that you can use as scalar inputs for field-MOP creation or evaluation.

  • To create a field-MOP, select Field data models Field-MOP Create Field-MOP.

  • To evaluate a field-MOP, select Field data models > Approximate from imported data > Field-MOP from inputs.

Once a field-MOP is created or evaluated, you can verify random field approximation.

Verifying Random Field Approximation

If the random field model's total explained variation is below 100%, the reduction from field design to random field amplitudes is an approximation. Consequently, you should verify the accuracy of the reverse conversion, from random field amplitudes to field design.

To verify random field approximation:

  1. Select all random field amplitudes created previously.

  2. Select Field data models > Approximate from imported data > Random field from amplitudes.

    oSP3D adds the following field data objects to the database:

    • Imported field designs, for which random field amplitudes have been calculated

    • Approximated field designs, which were calculated during the reverse conversion

  3. Select individual field data objects for direct comparison and then select Statistics > Errors and differences > Compute relative error.