Recognition of partial discharges using an ensemble of neural networks

A. Abubakar Mas'ud, Brian Stewart, S.G. McMeekin, A. Nesbitt

Research output: Chapter in Book/Report/Conference proceedingConference contribution book

3 Citations (Scopus)

Abstract

This paper introduces an improved method for classifying Partial Discharge (PD) patterns using Ensemble Neural Network (ENN) learning. The method is based on training several Neural Network (NN) models and combining their predictions. In this paper it is applied to the recognition of PD from artificially created poly-ethylene-terephthalate (PET) voids and in particular the ability of the ENN to categorise statistical Φ-q-n patterns for two different void sizes over 50 and 250 power cycles. The training data for the ENN comprises statistical parameters obtained from voids of 0.6mm and 1mm diameter. Measurements were made on three separately manufactured void samples for both these diameters. Similarities between the different PD measurements and different cycle captures is investigated using both a Single Neural Network (SNN) and the ENN. For each set of 3 void samples, each NN was trained and tested from the data of one PD void defect. Each NN was then tested using data from two other void geometries in order to determine the recognition abilities of both the ENN and SNN. The results show that the ENN always produces higher recognition efficiency for unseen data when compared to the SNN. It is also shown that ENN produces similar recognition predictions for PD patterns captured using either 50 or 250 power cycles while the SNN shows more sensitivity to the number of power cycles captured.
LanguageEnglish
Title of host publicationProceedings of the 2011 IEEE Conference on Electrical Insulation and Dielectric Phenomena (CEIDP 2011)
Place of PublicationPiscataway, N.J.
PublisherIEEE
Pages497-500
Number of pages4
ISBN (Print)9781457709852
DOIs
Publication statusPublished - 19 Oct 2011
Externally publishedYes

Fingerprint

Partial discharges
Neural networks
Polyethylene terephthalates

Keywords

  • partial discharge
  • pattern classification
  • neural networks

Cite this

Abubakar Mas'ud, A., Stewart, B., McMeekin, S. G., & Nesbitt, A. (2011). Recognition of partial discharges using an ensemble of neural networks. In Proceedings of the 2011 IEEE Conference on Electrical Insulation and Dielectric Phenomena (CEIDP 2011) (pp. 497-500). Piscataway, N.J.: IEEE. https://doi.org/10.1109/CEIDP.2011.6232703
Abubakar Mas'ud, A. ; Stewart, Brian ; McMeekin, S.G. ; Nesbitt, A. / Recognition of partial discharges using an ensemble of neural networks. Proceedings of the 2011 IEEE Conference on Electrical Insulation and Dielectric Phenomena (CEIDP 2011). Piscataway, N.J. : IEEE, 2011. pp. 497-500
@inproceedings{47cb54c86fa746ac81bef003b2462c76,
title = "Recognition of partial discharges using an ensemble of neural networks",
abstract = "This paper introduces an improved method for classifying Partial Discharge (PD) patterns using Ensemble Neural Network (ENN) learning. The method is based on training several Neural Network (NN) models and combining their predictions. In this paper it is applied to the recognition of PD from artificially created poly-ethylene-terephthalate (PET) voids and in particular the ability of the ENN to categorise statistical Φ-q-n patterns for two different void sizes over 50 and 250 power cycles. The training data for the ENN comprises statistical parameters obtained from voids of 0.6mm and 1mm diameter. Measurements were made on three separately manufactured void samples for both these diameters. Similarities between the different PD measurements and different cycle captures is investigated using both a Single Neural Network (SNN) and the ENN. For each set of 3 void samples, each NN was trained and tested from the data of one PD void defect. Each NN was then tested using data from two other void geometries in order to determine the recognition abilities of both the ENN and SNN. The results show that the ENN always produces higher recognition efficiency for unseen data when compared to the SNN. It is also shown that ENN produces similar recognition predictions for PD patterns captured using either 50 or 250 power cycles while the SNN shows more sensitivity to the number of power cycles captured.",
keywords = "partial discharge, pattern classification, neural networks",
author = "{Abubakar Mas'ud}, A. and Brian Stewart and S.G. McMeekin and A. Nesbitt",
year = "2011",
month = "10",
day = "19",
doi = "10.1109/CEIDP.2011.6232703",
language = "English",
isbn = "9781457709852",
pages = "497--500",
booktitle = "Proceedings of the 2011 IEEE Conference on Electrical Insulation and Dielectric Phenomena (CEIDP 2011)",
publisher = "IEEE",

}

Abubakar Mas'ud, A, Stewart, B, McMeekin, SG & Nesbitt, A 2011, Recognition of partial discharges using an ensemble of neural networks. in Proceedings of the 2011 IEEE Conference on Electrical Insulation and Dielectric Phenomena (CEIDP 2011). IEEE, Piscataway, N.J., pp. 497-500. https://doi.org/10.1109/CEIDP.2011.6232703

Recognition of partial discharges using an ensemble of neural networks. / Abubakar Mas'ud, A.; Stewart, Brian; McMeekin, S.G.; Nesbitt, A.

Proceedings of the 2011 IEEE Conference on Electrical Insulation and Dielectric Phenomena (CEIDP 2011). Piscataway, N.J. : IEEE, 2011. p. 497-500.

Research output: Chapter in Book/Report/Conference proceedingConference contribution book

TY - GEN

T1 - Recognition of partial discharges using an ensemble of neural networks

AU - Abubakar Mas'ud, A.

AU - Stewart, Brian

AU - McMeekin, S.G.

AU - Nesbitt, A.

PY - 2011/10/19

Y1 - 2011/10/19

N2 - This paper introduces an improved method for classifying Partial Discharge (PD) patterns using Ensemble Neural Network (ENN) learning. The method is based on training several Neural Network (NN) models and combining their predictions. In this paper it is applied to the recognition of PD from artificially created poly-ethylene-terephthalate (PET) voids and in particular the ability of the ENN to categorise statistical Φ-q-n patterns for two different void sizes over 50 and 250 power cycles. The training data for the ENN comprises statistical parameters obtained from voids of 0.6mm and 1mm diameter. Measurements were made on three separately manufactured void samples for both these diameters. Similarities between the different PD measurements and different cycle captures is investigated using both a Single Neural Network (SNN) and the ENN. For each set of 3 void samples, each NN was trained and tested from the data of one PD void defect. Each NN was then tested using data from two other void geometries in order to determine the recognition abilities of both the ENN and SNN. The results show that the ENN always produces higher recognition efficiency for unseen data when compared to the SNN. It is also shown that ENN produces similar recognition predictions for PD patterns captured using either 50 or 250 power cycles while the SNN shows more sensitivity to the number of power cycles captured.

AB - This paper introduces an improved method for classifying Partial Discharge (PD) patterns using Ensemble Neural Network (ENN) learning. The method is based on training several Neural Network (NN) models and combining their predictions. In this paper it is applied to the recognition of PD from artificially created poly-ethylene-terephthalate (PET) voids and in particular the ability of the ENN to categorise statistical Φ-q-n patterns for two different void sizes over 50 and 250 power cycles. The training data for the ENN comprises statistical parameters obtained from voids of 0.6mm and 1mm diameter. Measurements were made on three separately manufactured void samples for both these diameters. Similarities between the different PD measurements and different cycle captures is investigated using both a Single Neural Network (SNN) and the ENN. For each set of 3 void samples, each NN was trained and tested from the data of one PD void defect. Each NN was then tested using data from two other void geometries in order to determine the recognition abilities of both the ENN and SNN. The results show that the ENN always produces higher recognition efficiency for unseen data when compared to the SNN. It is also shown that ENN produces similar recognition predictions for PD patterns captured using either 50 or 250 power cycles while the SNN shows more sensitivity to the number of power cycles captured.

KW - partial discharge

KW - pattern classification

KW - neural networks

U2 - 10.1109/CEIDP.2011.6232703

DO - 10.1109/CEIDP.2011.6232703

M3 - Conference contribution book

SN - 9781457709852

SP - 497

EP - 500

BT - Proceedings of the 2011 IEEE Conference on Electrical Insulation and Dielectric Phenomena (CEIDP 2011)

PB - IEEE

CY - Piscataway, N.J.

ER -

Abubakar Mas'ud A, Stewart B, McMeekin SG, Nesbitt A. Recognition of partial discharges using an ensemble of neural networks. In Proceedings of the 2011 IEEE Conference on Electrical Insulation and Dielectric Phenomena (CEIDP 2011). Piscataway, N.J.: IEEE. 2011. p. 497-500 https://doi.org/10.1109/CEIDP.2011.6232703