An Adaptive Frequency Sweep Algorithm Employing Additive Perturbation in the Loewner Matrix Model for Electromagnetic Simulations

Abstract

In this paper, we have proposed a novel adaptive frequency sweep algorithm for electromagnetic simulation based on additive perturbations of the Loewner-State matrix. This algorithm incorporates the Loewner matrix rational interpolation method to model the frequency domain data. Two Loewner matrix models are generated from the simulated data, the first one from the basic Loewner matrix model and the second one by introducing an additive perturbation to the same model. The error between these two models is evaluated, and the frequency point corresponding to the maximum error is selected as the next simulation point. At each iteration, new frequency points are adaptively determined based on the proposed pseudo-error criterion. Two numerical examples are conducted to verify the accuracy and speed of this algorithm. The results show that the proposed Loewner matrix-based adaptive frequency sweep approach demonstrates exceptional performance in both speed and accuracy.