An investigative study into the sensitivity of different partial discharge phi-q-n pattern resolution sizes on statistical neural network pattern classification

Abdullahi Abubakar Mas Ud, Brian Stewart, Scott McMeekin

Research output: Contribution to journalArticle

20 Downloads (Pure)

Abstract

This paper investigates the sensitivity of statistical fingerprints to different phase resolution (PR) and amplitude bins (AB) sizes of partial discharge (PD) phi-q-n (phase-amplitude-number) patterns. In particular, this paper compares the capability of the nsemble neural network (ENN) and the single neural network (SNN) in recognizing and distinguishing different resolution sizes of phi-q-n discharge patterns. The training fingerprints for both the SNN and ENN comprise statistical fingerprints from different f-q-n measurements. The result shows that there exists statistical distinction for different PR and AB sizes on some of the statistical fingerprints. Additionally, the ENN and SNN outputs change depending on training and testing with different PR and AB sizes. Furthermore, the ENN appears to be more sensitive in recognizing and discriminating the resolution changes when compared with the SNN. Finally, the results are assessed for practical implementation in the power industry and benefits to practitioners in the field are highlighted.
Original languageEnglish
Pages (from-to)497–507
Number of pages11
JournalMeasurement
Volume92
Early online date23 Jun 2016
DOIs
Publication statusPublished - Oct 2016

Fingerprint

Partial discharges
Pattern recognition
Neural networks
Bins
sensitivity
education
industries
output
Testing
Industry

Keywords

  • classification
  • partial discharge
  • ensemble neural network
  • phase resolution and amplitude bin sizes

Cite this

@article{ac9588abdd7d4310a88007007aab2fde,
title = "An investigative study into the sensitivity of different partial discharge phi-q-n pattern resolution sizes on statistical neural network pattern classification",
abstract = "This paper investigates the sensitivity of statistical fingerprints to different phase resolution (PR) and amplitude bins (AB) sizes of partial discharge (PD) phi-q-n (phase-amplitude-number) patterns. In particular, this paper compares the capability of the nsemble neural network (ENN) and the single neural network (SNN) in recognizing and distinguishing different resolution sizes of phi-q-n discharge patterns. The training fingerprints for both the SNN and ENN comprise statistical fingerprints from different f-q-n measurements. The result shows that there exists statistical distinction for different PR and AB sizes on some of the statistical fingerprints. Additionally, the ENN and SNN outputs change depending on training and testing with different PR and AB sizes. Furthermore, the ENN appears to be more sensitive in recognizing and discriminating the resolution changes when compared with the SNN. Finally, the results are assessed for practical implementation in the power industry and benefits to practitioners in the field are highlighted.",
keywords = "classification, partial discharge, ensemble neural network, phase resolution and amplitude bin sizes",
author = "{Mas Ud}, {Abdullahi Abubakar} and Brian Stewart and Scott McMeekin",
note = "AAM uploaded 5-7-16 (Word file, re-uploaded PDF 5-8-16); 12m embargo accepted: 22/6/16 (in SAN) online pub: 23/7/16 Record was set to backend visibility by BS. Co-author SM confirmed can be made public. ET 12-9-16",
year = "2016",
month = "10",
doi = "10.1016/j.measurement.2016.06.043",
language = "English",
volume = "92",
pages = "497–507",
journal = "Measurement",
issn = "0263-2241",
publisher = "Elsevier B.V.",

}

An investigative study into the sensitivity of different partial discharge phi-q-n pattern resolution sizes on statistical neural network pattern classification. / Mas Ud, Abdullahi Abubakar; Stewart, Brian; McMeekin, Scott.

In: Measurement, Vol. 92, 10.2016, p. 497–507.

Research output: Contribution to journalArticle

TY - JOUR

T1 - An investigative study into the sensitivity of different partial discharge phi-q-n pattern resolution sizes on statistical neural network pattern classification

AU - Mas Ud, Abdullahi Abubakar

AU - Stewart, Brian

AU - McMeekin, Scott

N1 - AAM uploaded 5-7-16 (Word file, re-uploaded PDF 5-8-16); 12m embargo accepted: 22/6/16 (in SAN) online pub: 23/7/16 Record was set to backend visibility by BS. Co-author SM confirmed can be made public. ET 12-9-16

PY - 2016/10

Y1 - 2016/10

N2 - This paper investigates the sensitivity of statistical fingerprints to different phase resolution (PR) and amplitude bins (AB) sizes of partial discharge (PD) phi-q-n (phase-amplitude-number) patterns. In particular, this paper compares the capability of the nsemble neural network (ENN) and the single neural network (SNN) in recognizing and distinguishing different resolution sizes of phi-q-n discharge patterns. The training fingerprints for both the SNN and ENN comprise statistical fingerprints from different f-q-n measurements. The result shows that there exists statistical distinction for different PR and AB sizes on some of the statistical fingerprints. Additionally, the ENN and SNN outputs change depending on training and testing with different PR and AB sizes. Furthermore, the ENN appears to be more sensitive in recognizing and discriminating the resolution changes when compared with the SNN. Finally, the results are assessed for practical implementation in the power industry and benefits to practitioners in the field are highlighted.

AB - This paper investigates the sensitivity of statistical fingerprints to different phase resolution (PR) and amplitude bins (AB) sizes of partial discharge (PD) phi-q-n (phase-amplitude-number) patterns. In particular, this paper compares the capability of the nsemble neural network (ENN) and the single neural network (SNN) in recognizing and distinguishing different resolution sizes of phi-q-n discharge patterns. The training fingerprints for both the SNN and ENN comprise statistical fingerprints from different f-q-n measurements. The result shows that there exists statistical distinction for different PR and AB sizes on some of the statistical fingerprints. Additionally, the ENN and SNN outputs change depending on training and testing with different PR and AB sizes. Furthermore, the ENN appears to be more sensitive in recognizing and discriminating the resolution changes when compared with the SNN. Finally, the results are assessed for practical implementation in the power industry and benefits to practitioners in the field are highlighted.

KW - classification

KW - partial discharge

KW - ensemble neural network

KW - phase resolution and amplitude bin sizes

U2 - 10.1016/j.measurement.2016.06.043

DO - 10.1016/j.measurement.2016.06.043

M3 - Article

VL - 92

SP - 497

EP - 507

JO - Measurement

JF - Measurement

SN - 0263-2241

ER -