Selecting appropriate machine learning classifiers for DGA diagnosis

Jose Ignacio Aizpurua Unanue; Victoria Catterson; Brian G Stewart; Stephen McArthur; Brandon Lambert; Bismark Ampofo; Gavin  Pereira; James Cross

doi:10.1109/CEIDP.2017.8257475

Selecting appropriate machine learning classifiers for DGA diagnosis

Jose Ignacio Aizpurua Unanue, Victoria Catterson, Brian G Stewart, Stephen McArthur, Brandon Lambert, Bismark Ampofo, Gavin Pereira, James Cross

Electronic And Electrical Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution book

6 Citations (Scopus)

479 Downloads (Pure)

Abstract

Dissolved gas analysis (DGA) is a common method of assessing transformer health. There are a number of machine learning classifiers reported to give a high accuracy on specific datasets, such as Artificial Neural Networks or Support Vector Machines. When these methods reach the same conclusion about the type of fault present, this can give an increased confidence in the veracity of the diagnosis. However, it is critical to analyze and quantify the strength of these classifiers in the presence of conflicting data to test their practicality for usage in the field. This paper investigates the adequacy of different machine
learning based DGA diagnosis models in the presence of conflicting data. The proposed method will aid engineers with the selection of machine learning models so as to maximize the usability and accuracy in the presence of conflicting data.

Original language	English
Title of host publication	2017 IEEE Conference on Electrical Insulation and Dielectric Phenomena
Place of Publication	Piscataway, N.J.
Pages	153-156
Number of pages	4
DOIs	https://doi.org/10.1109/CEIDP.2017.8257475
Publication status	Published - 15 Jan 2018

Keywords

dissolved gas analysis
transformer health assessment
transformers

Access to Document

10.1109/CEIDP.2017.8257475

Aizpurua-etal-IEEE-CEIDP-2017-Selecting-appropriate-machine-learning-clasifiersAccepted author manuscript, 497 KB

Cite this

@inproceedings{7b9d74845eae45848dfba175f75204ea,

title = "Selecting appropriate machine learning classifiers for DGA diagnosis",

abstract = "Dissolved gas analysis (DGA) is a common method of assessing transformer health. There are a number of machine learning classifiers reported to give a high accuracy on specific datasets, such as Artificial Neural Networks or Support Vector Machines. When these methods reach the same conclusion about the type of fault present, this can give an increased confidence in the veracity of the diagnosis. However, it is critical to analyze and quantify the strength of these classifiers in the presence of conflicting data to test their practicality for usage in the field. This paper investigates the adequacy of different machinelearning based DGA diagnosis models in the presence of conflicting data. The proposed method will aid engineers with the selection of machine learning models so as to maximize the usability and accuracy in the presence of conflicting data.",

keywords = "dissolved gas analysis, transformer health assessment, transformers",

author = "{Aizpurua Unanue}, {Jose Ignacio} and Victoria Catterson and Stewart, {Brian G} and Stephen McArthur and Brandon Lambert and Bismark Ampofo and Gavin Pereira and James Cross",

note = "{\textcopyright} 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting /republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.",

year = "2018",

month = jan,

day = "15",

doi = "10.1109/CEIDP.2017.8257475",

language = "English",

isbn = "978-1-5386-1194-4",

pages = "153--156",

booktitle = "2017 IEEE Conference on Electrical Insulation and Dielectric Phenomena",

}

TY - GEN

T1 - Selecting appropriate machine learning classifiers for DGA diagnosis

AU - Aizpurua Unanue, Jose Ignacio

AU - Catterson, Victoria

AU - Stewart, Brian G

AU - McArthur, Stephen

AU - Lambert, Brandon

AU - Ampofo, Bismark

AU - Pereira, Gavin

AU - Cross, James

N1 - © 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting /republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

PY - 2018/1/15

Y1 - 2018/1/15

N2 - Dissolved gas analysis (DGA) is a common method of assessing transformer health. There are a number of machine learning classifiers reported to give a high accuracy on specific datasets, such as Artificial Neural Networks or Support Vector Machines. When these methods reach the same conclusion about the type of fault present, this can give an increased confidence in the veracity of the diagnosis. However, it is critical to analyze and quantify the strength of these classifiers in the presence of conflicting data to test their practicality for usage in the field. This paper investigates the adequacy of different machinelearning based DGA diagnosis models in the presence of conflicting data. The proposed method will aid engineers with the selection of machine learning models so as to maximize the usability and accuracy in the presence of conflicting data.

AB - Dissolved gas analysis (DGA) is a common method of assessing transformer health. There are a number of machine learning classifiers reported to give a high accuracy on specific datasets, such as Artificial Neural Networks or Support Vector Machines. When these methods reach the same conclusion about the type of fault present, this can give an increased confidence in the veracity of the diagnosis. However, it is critical to analyze and quantify the strength of these classifiers in the presence of conflicting data to test their practicality for usage in the field. This paper investigates the adequacy of different machinelearning based DGA diagnosis models in the presence of conflicting data. The proposed method will aid engineers with the selection of machine learning models so as to maximize the usability and accuracy in the presence of conflicting data.

KW - dissolved gas analysis

KW - transformer health assessment

KW - transformers

UR - https://www.ieee.org/conferences_events/conferences/conferencedetails/index.html

U2 - 10.1109/CEIDP.2017.8257475

DO - 10.1109/CEIDP.2017.8257475

M3 - Conference contribution book

SN - 978-1-5386-1194-4

SP - 153

EP - 156

BT - 2017 IEEE Conference on Electrical Insulation and Dielectric Phenomena

CY - Piscataway, N.J.

ER -

Selecting appropriate machine learning classifiers for DGA diagnosis

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Prognostics & health management oriented data analytics suite for transformer health monitoring

Cite this