Design of Neural Network Based Automatic Voltage Regulator for Synchronous Generator

Abstract

variation in load on the system, which damages the equipment of the power system. In this paper a PID controller and a Neural Network for Automatic Voltage Regulator (AVR) has been designed in order to reduce these voltage fluctuations and to provide proper stability to the system. The neural network has been applied to SMIB system model with the objective of improving the performance. The parameters of the neural network configuration have been provided so that the output of the system is efficient enough to prevent the perturbations irrespective of the changes in mechanical torque and also to maintain the voltage level that has been provided as the input. Another controller namely PID controller is also used to the AVR. The PID constant values are tuned by PID real-time auto-tuning in MathWorks which ensures closed loop stability, adequate performance and thus, providing efficient performance. The PID constants values are further changed for generating the output closer to the set value. The simulation was done in MATLAB SIMULINK and the results have been discussed.