Application of an ensemble neural network for classifying partial discharge patterns

A.A.a Mas'ud, B.G. Stewart, S.G. McMeekin

Research output: Contribution to journalArticle

24 Citations (Scopus)

Abstract

This paper proposes a technique for classifying partial discharge (PD) patterns based on ensemble neural network (ENN) learning. The ENN technique is based on training a number of neural network (NN) models with statistical parameters from PD patterns and combining their predictions. In this paper, six constituent NN models form the ensemble. Combining the outputs of the constituent NNs through an aggregating unit using dynamically weighted averaging strategy gives a final evaluation of PD patterns in relation to a range of PD fault types. Using the data sets of measured PD patterns as the system input fingerprints, the classification performance of the ENN has been compared statistically and quantitatively with a single neural network (SNN). This is achieved through evaluating the average, variance and standard error of the means of ENN and SNN recognition performances over 100 different initial states of the NNs thus providing an effective comparison to be made. The result shows that the ENN appears to be more robust with statistically improved performance in recognizing untrained PD patterns for a number of PD fault geometries. © 2014 Elsevier B.V.
LanguageEnglish
Pages154-162
Number of pages9
JournalElectric Power Systems Research
Volume110
DOIs
Publication statusPublished - 31 May 2014
Externally publishedYes

Fingerprint

Partial discharges
Neural networks
Geometry

Keywords

  • partial discharge
  • ensemble neural network
  • single neural network

Cite this

@article{af0b48f1951a4bd397ebbe919e62a7c5,
title = "Application of an ensemble neural network for classifying partial discharge patterns",
abstract = "This paper proposes a technique for classifying partial discharge (PD) patterns based on ensemble neural network (ENN) learning. The ENN technique is based on training a number of neural network (NN) models with statistical parameters from PD patterns and combining their predictions. In this paper, six constituent NN models form the ensemble. Combining the outputs of the constituent NNs through an aggregating unit using dynamically weighted averaging strategy gives a final evaluation of PD patterns in relation to a range of PD fault types. Using the data sets of measured PD patterns as the system input fingerprints, the classification performance of the ENN has been compared statistically and quantitatively with a single neural network (SNN). This is achieved through evaluating the average, variance and standard error of the means of ENN and SNN recognition performances over 100 different initial states of the NNs thus providing an effective comparison to be made. The result shows that the ENN appears to be more robust with statistically improved performance in recognizing untrained PD patterns for a number of PD fault geometries. {\circledC} 2014 Elsevier B.V.",
keywords = "partial discharge, ensemble neural network, single neural network",
author = "A.A.a Mas'ud and B.G. Stewart and S.G. McMeekin",
year = "2014",
month = "5",
day = "31",
doi = "10.1016/j.epsr.2014.01.010",
language = "English",
volume = "110",
pages = "154--162",
journal = "Electric Power Systems Research",
issn = "0378-7796",

}

Application of an ensemble neural network for classifying partial discharge patterns. / Mas'ud, A.A.a; Stewart, B.G.; McMeekin, S.G.

In: Electric Power Systems Research, Vol. 110, 31.05.2014, p. 154-162.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Application of an ensemble neural network for classifying partial discharge patterns

AU - Mas'ud, A.A.a

AU - Stewart, B.G.

AU - McMeekin, S.G.

PY - 2014/5/31

Y1 - 2014/5/31

N2 - This paper proposes a technique for classifying partial discharge (PD) patterns based on ensemble neural network (ENN) learning. The ENN technique is based on training a number of neural network (NN) models with statistical parameters from PD patterns and combining their predictions. In this paper, six constituent NN models form the ensemble. Combining the outputs of the constituent NNs through an aggregating unit using dynamically weighted averaging strategy gives a final evaluation of PD patterns in relation to a range of PD fault types. Using the data sets of measured PD patterns as the system input fingerprints, the classification performance of the ENN has been compared statistically and quantitatively with a single neural network (SNN). This is achieved through evaluating the average, variance and standard error of the means of ENN and SNN recognition performances over 100 different initial states of the NNs thus providing an effective comparison to be made. The result shows that the ENN appears to be more robust with statistically improved performance in recognizing untrained PD patterns for a number of PD fault geometries. © 2014 Elsevier B.V.

AB - This paper proposes a technique for classifying partial discharge (PD) patterns based on ensemble neural network (ENN) learning. The ENN technique is based on training a number of neural network (NN) models with statistical parameters from PD patterns and combining their predictions. In this paper, six constituent NN models form the ensemble. Combining the outputs of the constituent NNs through an aggregating unit using dynamically weighted averaging strategy gives a final evaluation of PD patterns in relation to a range of PD fault types. Using the data sets of measured PD patterns as the system input fingerprints, the classification performance of the ENN has been compared statistically and quantitatively with a single neural network (SNN). This is achieved through evaluating the average, variance and standard error of the means of ENN and SNN recognition performances over 100 different initial states of the NNs thus providing an effective comparison to be made. The result shows that the ENN appears to be more robust with statistically improved performance in recognizing untrained PD patterns for a number of PD fault geometries. © 2014 Elsevier B.V.

KW - partial discharge

KW - ensemble neural network

KW - single neural network

UR - http://www.sciencedirect.com/science/article/pii/S0378779614000145

UR - https://www.scopus.com/inward/record.uri?eid=2-s2.0-84894252201&partnerID=40&md5=a601d640abea7ce678104c320e57315c

U2 - 10.1016/j.epsr.2014.01.010

DO - 10.1016/j.epsr.2014.01.010

M3 - Article

VL - 110

SP - 154

EP - 162

JO - Electric Power Systems Research

T2 - Electric Power Systems Research

JF - Electric Power Systems Research

SN - 0378-7796

ER -