Stability Enhancement of Synchronous Generators Using ANN-Based Power System Stabilizers '

Abstract

This thesis presents a theoretical and experimental study of the petformance of a synchronous generator connected to an infinite bus via two parallel transmission lines, when it is equipped with an automatic voltage regulator and a power system stabilizer (PSS). The excitation loop together with other factors such as weak transmission, and load characteristics may lead to negative damping into the network. As a result, minor disturbances could set off sustained oscillations. PSS has been proven as an effective way of damping system oscillations. The PSS suggested in this thesis depends on the artificial neural network technique.

The • thesis • presents a mathematical • model for the synchronous . generator equipped once with a lead-lag PSS, and then with ANN based PSS. The lead-lag PSS is used for two main purposes, firstly to train the ANN network, secondly it has been experimentally implemented to validate the simulation results further obtained when the machine should be equipped with ANN-based PSS.

An artificial •neural network (ANN) trained to function as a power system stabilizer. (PSS) is presented. In order to make the proposed ANN PSS work properly, it was trained over the full working range of tl1e generating llllit witl1 a large variety of disturbances. Data used to train the ANN PSS are collected from two main sources. The first is tl1e output of a welL tlllled lead-lag PSS fitted to the machine. The second source of data is the synchronous machine •deviation from its nominal value.