Contents I. Introduction
The finite-element (FE) method is commonly used to study low-frequency electromagnetic devices. This approach gives accurate results but requires large computational times due to numerical or physical features such as a high number of degrees of freedom (DoFs) in space and also a high number of time steps or the nonlinear behavior of ferromagnetic materials for example. In order to reduce the computational time, especially for the parametrized model, model order reduction methods have been proposed in the literature. One of the most popular approaches is the proper orthogonal decomposition (POD) approach. Based on the solutions of the FE model for different values of parameters (called snapshots), the POD enables to approximate the solution of the FE model in a reduced basis [1]. Then, the initial FE system is projected onto a reduced basis, decreasing the order of the numerical model to be solved for new parameter values. Another approach consists in constructing a metamodel to interpolate directly the solution expressed into a reduced basis for new parameter values. Different approaches can be used, as for example, based on an optimization process [2] or polynomial functions [3]. The radial basis function (RBF) interpolation method can be also applied in this context. In the literature, the POD-RBF approach has been developed for mechanical or thermal problems (see [4], [5]).
