A Summary of Artificial Neural Networks on Electromagnetic Interference Diagnosis

摘要

Artificial neural networks (ANN) have achieved successes in many fields, such as pattern recognition and speech recognition etc.. In this paper, we will summary its applications in the source reconstruction for electromagnetic interference (EMI) diagnosis in two ways. Firstly, near-field scanning is used to obtain the near-field data, and then ANN is used to reconstruct the equivalent EMI sources from the near-field. Both amplitude and phase of the near-field are used in our methods. The first case is to show the ability of ANN to consider multi-reflection and/or diffraction effects in EMI source reconstruction. The next case is to show how to reconstruct EMI sources in an iterative way with ANN. Both two cases are verified by measurement examples.