Data-driven structured identification of 2D Maxwell’s equation as a port-Hamiltonian system

The objective of this work is to provide a systematic procedure for Reduced Order Modelling of the 2D Maxwell’s equations with collocated boundary control and observation, on the basis of frequency domain data obtained thanks to a Mixed Finite Element Method: both the High Fidelity and the Low Fidelity Models share the so-called port-Hamiltonian structure (pHs) of the original PDE, which is preserved through the different steps. The efficient reduction technique relies on the Loewner framework, it is based on Benner et al. contribution (2020), which has been recently adapted to handle non-strictly passive model, and numerical issues observed when dealing with complex configurations.

To cite this paper:
Gouzien Mattéo, Poussot-Vassal Charles, Haine Ghislain, Matignon Denis (2024) Data-driven structured identification of 2D Maxwell's equation as a port-Hamiltonian system. In the proceedings of the 10ème Conférence Européenne sur les Méthodes Numériques en Electromagnétisme. Toulouse, France.

        Loading BibTeX…