Unsupervised Ann Based Pid Controller for a Superconducting Generator in a Multi-Machine Power System

Abstract

The paper presents an artificial neural network (ANN) based Proportional Integral Derivative (PID) controller for a superconducting generator (SCG) in a multi-machine power system. The studied power system includes a SCG and three conventional machines of different types and ratings. The SCG is controlled also by conventional PID, which is designed according to pole placement technique, implemented in its governor control loop. While, conventional generating units are controlled by different conventional excitation control systems. The ANN controller patterns are gathered from a simple unsupervised (self learning) ANN-PID using an optimization technique. To achieve a high degree of accuracy, the system is represented by a fairly detailed non-linear model. The simulation results reveal that, the proposed ANN controller is achieving further enhancement of the system performance in terms of damping increase and fast return of system variables to their nominal values over a wide range of operating conditions and under sever disturbances such as 3-phase short circuit, step increase in load, and 3-phase short circuit followed by one line outage compared to conventional PID controller.