Model order reduction techniques for electromagnetic macromodelling based on finite methods

Model order reduction of an electromagnetic system is understood as the approximation of a continuous or discrete model of the system by one of substantially lower order, yet capable of capturing the electromagnetic behaviour of the original one with sufficient engineering accuracy. Model order reduction technqiues are reviewed and critically examined in this paper. The emphasis is on techniques suitable for the generation of high-order Padé approximations to transfer functions of electromagnetic systems discretized using finite methods. The computational complexity associated with the application of such model order reduction techniques to electromagnetic systems of practical interest is discussed, and the computationally most efficient model order reduction algorithm is identified. The benefits of model order reduction are demonstrated through a series of numerical examples from the analysis of electromagnetic waveguides.