Modeling of the Maximum Induced Currents in Automotive Radiated Immunity Tests via Thevenin-based Metamodels

This paper presents three different metamodels for the prediction of the maximum current induced on key vehicle electronic units during an automotive radiated immunity test. The proposed modeling approach is based on a Thevenin circuital interpretation of the test setup which is estimated from a small set of measurements or simulations. The FFT-based trigonometric regression, the support vector machine and the Gaussian process regression are then applied to provide three different metamodels able of predicting the spectrum of the induced currents for any value of the incidence angle of the external EM field. The accuracy and the convergence of the proposed alternatives are investigated by comparing model predictions with the results obtained by means of a parametric full-wave electromagnetic simulation.