Validation of Positive Fraction Vector Fitting Algorithm in the identification of Passive Immittances

The paper deals with the problem of identification of guaranteed passive rational models for linear distributes structures. The identification is pursued in two separate steps: in the first step the Vector Fitting algorithm is used for the identification of the poles; then the identification of residues is formulated as a convex programming problem with constraints on single terms of the pole-residue expansion. This approach, although suboptimal, is easier than constraining passivity on the whole transfer function, and at the same time more rigorous than perturbative approaches. In the paper the procedure is validated in reference cases and comparatively evaluated vs. other existing passivity enforcement schemes.