An Adaptive Multi-Point Multi-Moment Model Order Reduction Algorithm for Fast Broadband Simulation of Large-Scale 3D Electromagnetic Models

摘要

Modeling complex high frequency circuits in three dimensions poses a demand for fast computation of impedance or scattering parameters, especially for large scale systems with matrix dimensions up to 100 millions. Having a push button solution in mind, we developed an adaptive rational Krylov subspace based model order reduction algorithm that automatically selects expansion points and moments to be matched based on an error indicator. Employing a SiP-structure and two other real world examples, a benchmark section demonstrates that our method performs significantly better three other methods proposed in the literature.