Assessment of Human Exposure in Case of Wireless Power Transfer for Automotive Applications using Stochastic Models

In this paper, different non-intrusive stochastic approaches are compared in view of human exposure assessment from an inductive power transfer system at 85 kHz, dedicated to automotive applications. The stochastic approaches are combined with a 3D finite element method to build adequate meta-models based on Kriging and Polynomial Chaos Expansion. These models are used to consider the uncertainty and variability of several parameters defining the electromagnetic problem. Such fast predictions of uncertainties may help to improve the design of shields for inductive power transfer systems considering health and safety standards.