Passive reduced-order modeling of electromagnetic systems Original Research Article

Rapid electromagnetic field simulation is now being recognized as a critical component of next-generation computer-aided design frameworks that will be used for performance evaluation and design of the information processing and communication systems of the 21st century. Except for very simple systems, computer simulation of electromagnetic interactions is hindered by the very large number of degrees of freedom involved in the discrete model. One potentially useful approach to overcoming this computational bottleneck is model order reduction, where parts of the electromagnetic model are replaced by models of substantially lower order, yet capable of capturing the electromagnetic behavior of the original subsystems with sufficient engineering accuracy. Possible approaches to the development of such model order reduction techniques are the subject of this paper. The potential applications and the benefits from such electromagnetic model order reduction are discussed. The important issue of passivity of the generated reduced model is considered, and a set of constraints on the state representation of the discrete electromagnetic boundary-value problem are identified for the reduced order model to be passive. The paper concludes with a presentation of several numerical examples from the application of model order reduction techniques to the analysis of electromagnetic waveguides and antennas.