Deep complex neural network learning for high-voltage insulation fault classification from complex bispectrum representation

Imene Mitiche*, Mark Jenkins, Philip Boreham, Alan Nesbitt, Gordon Morison

*Corresponding author for this work

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

10 Downloads (Pure)

Abstract

Bispectrum representations previously achieved a successful classification of insulation fault signals in High-Voltage (HV) power plant. The magnitude information of the Bispectrum was implemented as a feature for a Deep Neural Network. This preliminary research brought interest in evaluating the performance of Bispectrum as complex input features that are implemented into a Deep Complex Valued Convolutional Neural Network (CV-CNN). This paper presents the application of this novel method to condition monitoring of High Voltage (HV) power plant equipment. Discharge signals related to HV insulation faults are measured in a real-world power plant using the Electromagnetic Interference (EMI) method and processed using third order Higher-Order Statistics (HOS) to obtain a Bispectrum representation. By mapping the time-domain signal to Bispectrum representations the problem can be approached as a complex-valued classification task. This allows for the novel combination of complex Bispectrum and CV-CNN applied to the classification of HV discharge signals. The network is trained on signals from 9 classes and achieves high classification accuracy in each category, improving upon the performance of a Real Valued CNN (RV-CNN).
Original languageEnglish
Title of host publicationEuropean Signal Processing Conference Proceedings
PublisherIEEE
Number of pages5
Publication statusAccepted/In press - 3 Jun 2019

Fingerprint

Insulation
Neural networks
Power plants
Electric potential
Higher order statistics
Condition monitoring
Signal interference

Cite this

@inproceedings{7db4b22dedb842c88f88c2a4e53cef88,
title = "Deep complex neural network learning for high-voltage insulation fault classification from complex bispectrum representation",
abstract = "Bispectrum representations previously achieved a successful classification of insulation fault signals in High-Voltage (HV) power plant. The magnitude information of the Bispectrum was implemented as a feature for a Deep Neural Network. This preliminary research brought interest in evaluating the performance of Bispectrum as complex input features that are implemented into a Deep Complex Valued Convolutional Neural Network (CV-CNN). This paper presents the application of this novel method to condition monitoring of High Voltage (HV) power plant equipment. Discharge signals related to HV insulation faults are measured in a real-world power plant using the Electromagnetic Interference (EMI) method and processed using third order Higher-Order Statistics (HOS) to obtain a Bispectrum representation. By mapping the time-domain signal to Bispectrum representations the problem can be approached as a complex-valued classification task. This allows for the novel combination of complex Bispectrum and CV-CNN applied to the classification of HV discharge signals. The network is trained on signals from 9 classes and achieves high classification accuracy in each category, improving upon the performance of a Real Valued CNN (RV-CNN).",
author = "Imene Mitiche and Mark Jenkins and Philip Boreham and Alan Nesbitt and Gordon Morison",
note = "Acceptance in SAN AAM: no embargo",
year = "2019",
month = "6",
day = "3",
language = "English",
booktitle = "European Signal Processing Conference Proceedings",
publisher = "IEEE",

}

Deep complex neural network learning for high-voltage insulation fault classification from complex bispectrum representation. / Mitiche, Imene; Jenkins, Mark; Boreham, Philip ; Nesbitt, Alan; Morison, Gordon.

European Signal Processing Conference Proceedings. IEEE, 2019.

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

TY - GEN

T1 - Deep complex neural network learning for high-voltage insulation fault classification from complex bispectrum representation

AU - Mitiche, Imene

AU - Jenkins, Mark

AU - Boreham, Philip

AU - Nesbitt, Alan

AU - Morison, Gordon

N1 - Acceptance in SAN AAM: no embargo

PY - 2019/6/3

Y1 - 2019/6/3

N2 - Bispectrum representations previously achieved a successful classification of insulation fault signals in High-Voltage (HV) power plant. The magnitude information of the Bispectrum was implemented as a feature for a Deep Neural Network. This preliminary research brought interest in evaluating the performance of Bispectrum as complex input features that are implemented into a Deep Complex Valued Convolutional Neural Network (CV-CNN). This paper presents the application of this novel method to condition monitoring of High Voltage (HV) power plant equipment. Discharge signals related to HV insulation faults are measured in a real-world power plant using the Electromagnetic Interference (EMI) method and processed using third order Higher-Order Statistics (HOS) to obtain a Bispectrum representation. By mapping the time-domain signal to Bispectrum representations the problem can be approached as a complex-valued classification task. This allows for the novel combination of complex Bispectrum and CV-CNN applied to the classification of HV discharge signals. The network is trained on signals from 9 classes and achieves high classification accuracy in each category, improving upon the performance of a Real Valued CNN (RV-CNN).

AB - Bispectrum representations previously achieved a successful classification of insulation fault signals in High-Voltage (HV) power plant. The magnitude information of the Bispectrum was implemented as a feature for a Deep Neural Network. This preliminary research brought interest in evaluating the performance of Bispectrum as complex input features that are implemented into a Deep Complex Valued Convolutional Neural Network (CV-CNN). This paper presents the application of this novel method to condition monitoring of High Voltage (HV) power plant equipment. Discharge signals related to HV insulation faults are measured in a real-world power plant using the Electromagnetic Interference (EMI) method and processed using third order Higher-Order Statistics (HOS) to obtain a Bispectrum representation. By mapping the time-domain signal to Bispectrum representations the problem can be approached as a complex-valued classification task. This allows for the novel combination of complex Bispectrum and CV-CNN applied to the classification of HV discharge signals. The network is trained on signals from 9 classes and achieves high classification accuracy in each category, improving upon the performance of a Real Valued CNN (RV-CNN).

M3 - Conference contribution

BT - European Signal Processing Conference Proceedings

PB - IEEE

ER -