Analysis of Three Port Circulator Using Artificial Neural Networks

Abstract

A Radio Frequency (RF) three port circulator is a non reciprocal passive device, where microwave signal entering in one port and transmitting into next port in rotation. Now a days these element become crucial in microwave systems and antenna networks in which the high power signals to be directed and isolated i.e., it protects the microwave components from unwanted reflections. An exact model is need for analysis of these circulators that can be obtained from Electro Magnetic (EM) Simulations, but the computational cost is very high. Therefore, the Artificial Neural Network model is the best alternative to obtain a model with less computational cost. ANN can solve large variety of Radio Frequency and CAD related problems. The neural network model is trained and tested by different ANN Algorithms and tabulated for their performance comparison. The circulator is designed using inverse model. In this paper proposed inverse modes and direct inverse models are compared.