Reduced-Order Mode Discovery in Arbitrary Cavities via Proper Orthogonal Decomposition

The Proper Orthogonal Decomposition (POD) is a technique that extracts a set of mutually orthogonal spatiotemporal modes from a dataset of interest; these modes can then be used for analysis of the underlying dynamics of the problem or to construct reduced-order models for fast simulation of new results. Here, we apply the POD to electromagnetic cavities; for canonical cases the technique is able to recover the cavity modes exactly, while for cavities of arbitrary shape or arbitrary filling we are able to extract the dominant (most energetic) modes in a simple and fast way with no modification to the algorithm.