Prediction of magnetic field emissions by current source reconstruction using radial basis function network

Near-field measurement has been adopted in the prediction of emission levels for electromagnetic compatibility diagnosis and design purposes. A novel approach for effective reconstruction of the equivalent current source based on the near-field measurement data is presented, for the prediction of magnetic field emissions elsewhere. The approach consists of two steps: first, the distribution of the magnetic field component normal to the entire measurement plane is acquired by interpolation of the discrete measurement data using a radial basis function network; secondly, the magnetic emissions elsewhere are evaluated from the equivalent current source derived from the acquired magnetic field distribution. This approach requires the measurement of only one single magnetic field component. The accuracy of the proposed approach has been demonstrated by comparing the predicted field to that of the full-wave simulation; the robustness of the approach against measurement noise has also been verified.