Neural network approach to fault diagnosis in a distribution system

Azah Mohamed, M. D A Mazumder

Research output: Contribution to journalArticle

18 Citations (Scopus)

Abstract

The application of artificial neural network (ANN) to fault diagnosis in a distribution system is presented. The function of fault diagnosis is to locate faults, identify the faulty devices, and isolate the faulty sections. Hierarchical distributed ANN is developed for determining the fault location whereas single stage ANN modules are considered for identifying the faulty devices and isolating the faulty sections of the distribution system. The ANN design has been developed by considering both the structure and functions of the distribution system. The ANN models are based on the feedforward network and the training uses the error back-propagation algorithm. In the generation of training sets, the training patterns are generated based on the logic operation of relays and circuit breakers corresponding to the respective fault location for cases of single and multiple faults. The proposed fault diagnosis system has been designed for a practical distribution system. The simulation results show that the proposed method using ANN can accurately locate faults, identify the faulty devices and isolate the faulty line and bus sections.

Original languageEnglish
Pages (from-to)129-134
Number of pages6
JournalInternational Journal of Power and Energy Systems
Volume19
Issue number2
Publication statusPublished - 1999
Externally publishedYes

Fingerprint

Distribution System
Fault Diagnosis
Failure analysis
Artificial Neural Network
Neural Networks
Neural networks
Fault
Electric fault location
Feedforward Networks
Error Propagation
Backpropagation algorithms
Back-propagation Algorithm
Electric circuit breakers
Network Design
Neural Network Model
Relay
Logic
Module
Line
Training

ASJC Scopus subject areas

  • Energy (miscellaneous)

Cite this

Neural network approach to fault diagnosis in a distribution system. / Mohamed, Azah; Mazumder, M. D A.

In: International Journal of Power and Energy Systems, Vol. 19, No. 2, 1999, p. 129-134.

Research output: Contribution to journalArticle

Mohamed, Azah ; Mazumder, M. D A. / Neural network approach to fault diagnosis in a distribution system. In: International Journal of Power and Energy Systems. 1999 ; Vol. 19, No. 2. pp. 129-134.
@article{ba67087823d043788f380c4c73fc6a25,
title = "Neural network approach to fault diagnosis in a distribution system",
abstract = "The application of artificial neural network (ANN) to fault diagnosis in a distribution system is presented. The function of fault diagnosis is to locate faults, identify the faulty devices, and isolate the faulty sections. Hierarchical distributed ANN is developed for determining the fault location whereas single stage ANN modules are considered for identifying the faulty devices and isolating the faulty sections of the distribution system. The ANN design has been developed by considering both the structure and functions of the distribution system. The ANN models are based on the feedforward network and the training uses the error back-propagation algorithm. In the generation of training sets, the training patterns are generated based on the logic operation of relays and circuit breakers corresponding to the respective fault location for cases of single and multiple faults. The proposed fault diagnosis system has been designed for a practical distribution system. The simulation results show that the proposed method using ANN can accurately locate faults, identify the faulty devices and isolate the faulty line and bus sections.",
author = "Azah Mohamed and Mazumder, {M. D A}",
year = "1999",
language = "English",
volume = "19",
pages = "129--134",
journal = "International Journal of Power and Energy Systems",
issn = "1078-3466",
publisher = "ACTA Press",
number = "2",

}

TY - JOUR

T1 - Neural network approach to fault diagnosis in a distribution system

AU - Mohamed, Azah

AU - Mazumder, M. D A

PY - 1999

Y1 - 1999

N2 - The application of artificial neural network (ANN) to fault diagnosis in a distribution system is presented. The function of fault diagnosis is to locate faults, identify the faulty devices, and isolate the faulty sections. Hierarchical distributed ANN is developed for determining the fault location whereas single stage ANN modules are considered for identifying the faulty devices and isolating the faulty sections of the distribution system. The ANN design has been developed by considering both the structure and functions of the distribution system. The ANN models are based on the feedforward network and the training uses the error back-propagation algorithm. In the generation of training sets, the training patterns are generated based on the logic operation of relays and circuit breakers corresponding to the respective fault location for cases of single and multiple faults. The proposed fault diagnosis system has been designed for a practical distribution system. The simulation results show that the proposed method using ANN can accurately locate faults, identify the faulty devices and isolate the faulty line and bus sections.

AB - The application of artificial neural network (ANN) to fault diagnosis in a distribution system is presented. The function of fault diagnosis is to locate faults, identify the faulty devices, and isolate the faulty sections. Hierarchical distributed ANN is developed for determining the fault location whereas single stage ANN modules are considered for identifying the faulty devices and isolating the faulty sections of the distribution system. The ANN design has been developed by considering both the structure and functions of the distribution system. The ANN models are based on the feedforward network and the training uses the error back-propagation algorithm. In the generation of training sets, the training patterns are generated based on the logic operation of relays and circuit breakers corresponding to the respective fault location for cases of single and multiple faults. The proposed fault diagnosis system has been designed for a practical distribution system. The simulation results show that the proposed method using ANN can accurately locate faults, identify the faulty devices and isolate the faulty line and bus sections.

UR - http://www.scopus.com/inward/record.url?scp=0032682259&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0032682259&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:0032682259

VL - 19

SP - 129

EP - 134

JO - International Journal of Power and Energy Systems

JF - International Journal of Power and Energy Systems

SN - 1078-3466

IS - 2

ER -