Equivalent circuit based neural network model of microstrip discontinuities

Abstract

This work is on equivalent circuit (EC) based artificial neural network (ANN) black-box model of the microstrip transmission line discontinuities. While inputs to the black-box are electrical properties of the substrate and the geometry of the transmission line together with all the discontinuities; outputs are the elements of the equivalent circuits and propagation properties of the line. ANN is trained and tested by means of the automated data generated by the programme using input-output relations. In this way, EC- ANN models are constructed for the open-end and gap discontinuities of the microstrip line and their performances are evaluated compared with the targets and the effects of these discontinuities are also given with their graphics.