Optimum Design for PID-ANN Controller for Automatic Voltage Regulator of Synchronous Generator

Abstract

Excitation control in the Synchronous Generators (SG) is one of the most important processes to achieve the constancy and stability of the terminal voltage. In conventional controllers, for a high degree of automatic voltage control, a high gain is required which may tend it to instability during facing large and sudden disturbances. In addition, the excitation system is a nonlinear in nature as a result of the variation of the system’s parameters due to heat rising and electromagnetic parameters. This paper proposes the Proportional Integral Derivative (PID) controller based on an Artificial Neural Network (ANN) as an intelligent non-linear controller for the Automatic Voltage Regulator (AVR) of the three-phase synchronous generator. The proposed PID-ANN controller is designed according to the discrete presentation of the PID controller. The controller’s parameters are tuned by using the Particle Swarm Optimization (PSO) technique. The overall system is simulated by using MATLAB/SIMULINK program in addition to the traditional PI and IP controllers as a comparison references. Simulation results show that the PID-ANN has better performance than Proportional Integral (PI) and Integral Proportional (IP) controllers in viewpoint transient response and robustness. The margin of robustness for PID-ANN controller are tested using different SG's, the test shows that the controller can control all SG's with an acceptable response.