Statistical Modeling of Disturbed Antennas Based on the Polynomial Chaos Expansion

A new methodology of statistical modeling of the far field (FF) radiated by antennas undergoing random disturbances is presented. First, the radiated FF is transformed into a parsimonious form using the spherical modes expansion method (SMEM); then, a surrogate model relating the parsimonious field with the input random parameters is constructed using the polynomial chaos expansion method (PCEM). The combination of the SMEM and PCEM allows developing a compact and precise model with a minimized experimental design cost. The obtained model is computationally costless for generating statistical samples of disturbed antennas easily usable as surrogate models in various types of analyses. In order to demonstrate its performance, the proposed methodology is validated with a deformable canonical antenna-a dipole undergoing three independent random deformations (stretching, bending, and torsion), deriving a compact and precise surrogate model.