Discussion Forum

If I understand you correctly, the calculation you are describing is more in the category of post-processing. To compute far-fields from an antenna or scatterer, Comsol Multiphysics (with RF module) computes the fields in the region around the source (or scatterer) first, using finite elements. The resulting fields are typically "near-fields" because the volume is relatively small, due to computational considerations. The far fields are then found in post-processing, by an Aperture Field (AF) integration method, referred to as the Stratton-Chu method. This math is documented in the Cosmol help system for the RF modules (look up "far-field" in the help system). So you have a number of options here. Since you already have the near-field data yourself (from your test data) you could more directly write your own MatLab routines to implement basically the same far-field algorithm as is being used in Comsol, just by following the same formulas. If you want to involve Comsol Multiphysics so as to work within the Comsol environment, then you can probably import your data. You'll have to include the effective aperture geometry that goes with those data, of course. Then, you can probably trick Comsol Multiphysics into going straight into post-processing (or nearly so)by executing some kind of dummy run that doesn't compute anything really new, but which lets you apply the far-field post processing to the imported fields. I haven't tried this personally. My guess is that if you can successfully import your field test data onto an equivalent aperture surface (i.e., with the same geometry) in Comsol Multiphysics, then the rest of the job will be the relatively easy part. Good luck. (Oh, I should add that you will need a lot of near-field data for this to work. If you only have a very coarse sampling of the fields, then your problem will be much more challenging and may not have any realizable solution.)