Design of a neural-fuzzy algorithm for power system protection (H.V. Transmission line)‎

Dissertant

Shahin, Adi Abd al-Husayn Muhammad

Thesis advisor

Ibrahim, Ali Abd Allah

University

Omdurman Islamic University

Faculty

Faculty of Engineering

Department

Department of Electrical and Electronics Engineering

University Country

Sudan

Degree

Ph.D.

Degree Date

2012

English Abstract

Artificial neural networks and fuzzy logic are used in this research to achieve a high speed, more reliable and high accuracy fault detection for a high voltage transmission line. Traditional digital distance relay designs use estimation algorithms to determine the phasors of currents and voltages.

Current phasors are affected by D.C decaying noise, this lead to slower response time or inaccurate decision of digital relay in some cases. This research presents a modified distance relaying algorithm that use phasors of voltage only that has been combined with adaptive resonance theory (ART) neural network to eliminate the use of current phasors.

The neural network is trained to recognize faults on basis of a specific relay characteristic. The modified algorithm eliminates the need of phasors estimation of current which is used as neural networks inputs to determine the appearance of fault from relay location.

The neural network has been trained using training patterns generated from general normalized MHO characteristics. Three main software stages are used in the design.

In the first stage, a generation of fault patterns using concentric circles which represents impedances of line for different types of fault.

In second stage, adaptive resonance theory neural network has been trained for these patterns to perform sets of clusters that hold characteristics of different types of faults.

In third stage, a high voltage parallel transmission line test system is simulated using power system computer aided design (PSCAD) program.

Different fault conditions including different fault resistances and different fault inception angles are applied and a voltage signal has been read from one end of transmission line to generate input patterns to test the designed algorithm by calculate the distance between these patterns and the clusters that are generated in stage two using fuzzy logic. The test results have been compared with test results of relay algorithm use Discrete Fourier Transform (DFT) for the same power system.

The test results show that the relay is able to detect faults in lesser time with high accuracy as compared to conventional digital relay algorithm that use Discrete Fourier Transform (DFT) as phasors estimator algorithm.

Main Subjects

Electronic engineering

Topics

• Electric power systems
• Neural networks(Computer science)

No. of Pages

151

Table of Contents

• APA
• MLA
• AMA

Language

English

Data Type

Arab Theses

Record ID

BIM-364335