Title

Supervisory Level Identifier for a Multimachine Power System Using Radial Basis Function Neural Networks

Abstract

A Wide Area Controller (WAC) could give commands to all or certain selected controllable devices in a geographic region or area of a power system, such as generator excitation control, governor and turbine control, FACTS devices, transformer tap changers, loads, circuit breakers, etc, in order to optimize selected quantities or responses during steady state or transient operation. Such a wide area could have say m inputs to it which could be used to control n outputs. The WAC will need some model of every component in the area, in order to figure out the response from any one of the m inputs to any one of the n outputs. This could be in the form of differential equations, which would be very complicated for a large system, or it could be a neural network based identifier of this (nxm) control system. This paper develops such a (nxm) identifier using Radial Basis Function Neural Networks (RBFN) and tests it on a three machine system which also contains a STATCOM. The method can readily be expanded to much larger systems.

Department(s)

Electrical and Computer Engineering

Sponsor(s)

Duke Power Company
National Science Foundation (U.S.)

Keywords and Phrases

Multimachine Power System; Radial Basis Function Neural Networks; Static Compensator; Supervisory Level Identification; Wide Area Control

Document Type

Article - Conference proceedings

Document Version

Citation

File Type

text

Language(s)

English

Rights

© 2005 International Conference on Power System Operations and Planning (ICPSOP), All rights reserved.

This document is currently not available here.


Share

 
COinS