Suitable Feedforward Artificial Neural Network Automatic Voltage Regulator for Excitation Control System

Abstract

The feedforward artificial neural network (FFANN) based automatic voltage regulator (AVR) controller for excitation system using Matlab/Simulink approach is proposed in this paper.The proposed AVR controller investigates and demonstrates the application of radial basis function (RBF) and multiplayer perceptron (MLP) architectures of FFANNs and compares the suitability of both architectures. The simulation results of suggested AVR controller not only show the encouraging responses for the application but also show the improvement in the transient responses of synchronous machine. The responses of developed RBF and MLP networks have also been compared with conventional proportional integral and derivative (PID) excitation controller of synchronous machine. Investigations and results prove that FFANN AVR controllers are very simple and accurate than conventional AVR and also enhance the stability performance of synchronous machine in an efficient manner. This research also suggests that RBF network is more simple, accurate, fast and robust controller than MLP architecture. Huge numbers of research papers have been written and published on the different types of excitation controller, but the proposed controller for excitation control system is relatively most simple and suitable for software demonstration and practical implementation.