Non intrusive Polynomial Chaos-based stochastic macromodeling of multiport systems

We present a novel technique to efficiently perform the variability analysis of electromagnetic systems. The proposed method calculates a Polynomial Chaos-based macromodel of the system transfer function that includes its statistical properties. The combination of a non-intrusive Polynomial Chaos approach with the Vector Fitting algorithm allows to describe the system variability features with accuracy and efficiency. The results of the variability analysis performed with the proposed method are verified by means of comparison with respect to the standard Monte Carlo analysis.