Statistical analysis of power microwave circuits via polynomial chaos

A circuit simulation methodology is presented that allows an efficient statistical assessment of microwave circuits using standard design software. The approach is based on the representation of circuit variables in terms of polynomial chaos expansions and the derivation of equivalent electrical models that are simulated to obtain the pertinent coefficients. In accordance with recent literature results, the equivalent circuit models are generated via a suitable point-matching technique. This also allows an efficient modeling of nonlinear devices. The proposed method is fully compatible with standard SPICE-type simulators and corresponding built-in device models. An application example concerning the statistical analysis of a power amplifier validates the procedure and concludes the paper.