Detection of the Defective Vias in SIW Circuits From Single/Array Probe(s) Data Using Source Reconstruction Method and Machine Learning

In this article, a new approach to detect the defective vias in substrate-integrated waveguide (SIW) structures is proposed. First, very near-field radiations of SIW structures are measured using either a single magnetic field probe or a fast electronically switched probe array. A source reconstruction method is utilized to calculate the equivalent electric and magnetic currents on the surface of the SIW structure under investigation. Thereafter, these equivalent sources are used to obtain the magnetic fields very close to the sample boards. A machine learning algorithm is used to distinguish the radiations that are due to the defective vias from those because of radiating parts of the circuit such as feed lines. The simulation and measurement results confirm the validity and accuracy of this high-resolution method.