An Approach to Reducing Input Parameter Volume for Fault Classifiers

Research output: Contribution to journalArticle

Abstract

As condition monitoring of systems continues to grow in both complexity and application, an overabundance of data is amassed. Computational capabilities are unable to keep abreast of the subsequent processing requirements. Thus, a means of establishing computable prognostic models to accurately reflect process condition, whilst alleviating computational burdens, is essential. This is achievable by restricting the amount of information input that is redundant to modelling algorithms. In this paper, a variable clustering approach is investigated to reorganise the harmonics of common diagnostic features in rotating machinery into a smaller number of heterogeneous groups that reflect conditions of the machine with minimal information redundancy. Naïve Bayes classifiers established using a reduced number of highly sensitive input parameters realised superior classification powers over higher dimensional classifiers, demonstrating the effectiveness of the proposed approach. Furthermore, generic parameter capabilities were evidenced through confirmatory factor analysis. Parameters with superior deterministic power were identified alongside complimentary, uncorrelated, variables. Particularly, variables with little explanatory capacity could be eliminated and lead to further variable reductions. Their information sustainability is also evaluated with Naïve Bayes classifiers, showing that successive classification rates are sufficiently high when the first few harmonics are used. Further gains were illustrated on compression of chosen envelope harmonic features. A Naïve Bayes classification model incorporating just two compressed input variables realised an 83.3% success rate, both an increase in classification rate and an immense improvement volume-wise on the former ten parameter model.
LanguageEnglish
Pages199-212
Number of pages14
JournalInternational Journal of Automation and Computing
Volume16
Issue number2
Early online date23 Jan 2019
DOIs
Publication statusPublished - Apr 2019

Fingerprint

Fault
Classifiers
Classifier
Naive Bayes Classifier
Harmonic
Rotating machinery
Condition monitoring
Rotating Machinery
Factor analysis
Confirmatory Factor Analysis
Condition Monitoring
Naive Bayes
Redundancy
Sustainability
Sustainable development
Envelope
Diagnostics
High-dimensional
Continue
Compression

Cite this

@article{67e87191236c4eb39c9f65ca3807b3aa,
title = "An Approach to Reducing Input Parameter Volume for Fault Classifiers",
abstract = "As condition monitoring of systems continues to grow in both complexity and application, an overabundance of data is amassed. Computational capabilities are unable to keep abreast of the subsequent processing requirements. Thus, a means of establishing computable prognostic models to accurately reflect process condition, whilst alleviating computational burdens, is essential. This is achievable by restricting the amount of information input that is redundant to modelling algorithms. In this paper, a variable clustering approach is investigated to reorganise the harmonics of common diagnostic features in rotating machinery into a smaller number of heterogeneous groups that reflect conditions of the machine with minimal information redundancy. Na{\"i}ve Bayes classifiers established using a reduced number of highly sensitive input parameters realised superior classification powers over higher dimensional classifiers, demonstrating the effectiveness of the proposed approach. Furthermore, generic parameter capabilities were evidenced through confirmatory factor analysis. Parameters with superior deterministic power were identified alongside complimentary, uncorrelated, variables. Particularly, variables with little explanatory capacity could be eliminated and lead to further variable reductions. Their information sustainability is also evaluated with Na{\"i}ve Bayes classifiers, showing that successive classification rates are sufficiently high when the first few harmonics are used. Further gains were illustrated on compression of chosen envelope harmonic features. A Na{\"i}ve Bayes classification model incorporating just two compressed input variables realised an 83.3{\%} success rate, both an increase in classification rate and an immense improvement volume-wise on the former ten parameter model.",
keywords = "Fault diagnosis, Classification, Variable clustering, Data compression, Big data",
author = "Ann Smith and Fengshou Gu and Andrew Ball",
year = "2019",
month = "4",
doi = "10.1007/s11633-018-1162-7",
language = "English",
volume = "16",
pages = "199--212",
journal = "International Journal of Automation and Computing",
issn = "1476-8186",
publisher = "Chinese Academy of Sciences",
number = "2",

}

TY - JOUR

T1 - An Approach to Reducing Input Parameter Volume for Fault Classifiers

AU - Smith, Ann

AU - Gu, Fengshou

AU - Ball, Andrew

PY - 2019/4

Y1 - 2019/4

N2 - As condition monitoring of systems continues to grow in both complexity and application, an overabundance of data is amassed. Computational capabilities are unable to keep abreast of the subsequent processing requirements. Thus, a means of establishing computable prognostic models to accurately reflect process condition, whilst alleviating computational burdens, is essential. This is achievable by restricting the amount of information input that is redundant to modelling algorithms. In this paper, a variable clustering approach is investigated to reorganise the harmonics of common diagnostic features in rotating machinery into a smaller number of heterogeneous groups that reflect conditions of the machine with minimal information redundancy. Naïve Bayes classifiers established using a reduced number of highly sensitive input parameters realised superior classification powers over higher dimensional classifiers, demonstrating the effectiveness of the proposed approach. Furthermore, generic parameter capabilities were evidenced through confirmatory factor analysis. Parameters with superior deterministic power were identified alongside complimentary, uncorrelated, variables. Particularly, variables with little explanatory capacity could be eliminated and lead to further variable reductions. Their information sustainability is also evaluated with Naïve Bayes classifiers, showing that successive classification rates are sufficiently high when the first few harmonics are used. Further gains were illustrated on compression of chosen envelope harmonic features. A Naïve Bayes classification model incorporating just two compressed input variables realised an 83.3% success rate, both an increase in classification rate and an immense improvement volume-wise on the former ten parameter model.

AB - As condition monitoring of systems continues to grow in both complexity and application, an overabundance of data is amassed. Computational capabilities are unable to keep abreast of the subsequent processing requirements. Thus, a means of establishing computable prognostic models to accurately reflect process condition, whilst alleviating computational burdens, is essential. This is achievable by restricting the amount of information input that is redundant to modelling algorithms. In this paper, a variable clustering approach is investigated to reorganise the harmonics of common diagnostic features in rotating machinery into a smaller number of heterogeneous groups that reflect conditions of the machine with minimal information redundancy. Naïve Bayes classifiers established using a reduced number of highly sensitive input parameters realised superior classification powers over higher dimensional classifiers, demonstrating the effectiveness of the proposed approach. Furthermore, generic parameter capabilities were evidenced through confirmatory factor analysis. Parameters with superior deterministic power were identified alongside complimentary, uncorrelated, variables. Particularly, variables with little explanatory capacity could be eliminated and lead to further variable reductions. Their information sustainability is also evaluated with Naïve Bayes classifiers, showing that successive classification rates are sufficiently high when the first few harmonics are used. Further gains were illustrated on compression of chosen envelope harmonic features. A Naïve Bayes classification model incorporating just two compressed input variables realised an 83.3% success rate, both an increase in classification rate and an immense improvement volume-wise on the former ten parameter model.

KW - Fault diagnosis

KW - Classification

KW - Variable clustering

KW - Data compression

KW - Big data

UR - http://www.scopus.com/inward/record.url?scp=85060597368&partnerID=8YFLogxK

U2 - 10.1007/s11633-018-1162-7

DO - 10.1007/s11633-018-1162-7

M3 - Article

VL - 16

SP - 199

EP - 212

JO - International Journal of Automation and Computing

T2 - International Journal of Automation and Computing

JF - International Journal of Automation and Computing

SN - 1476-8186

IS - 2

ER -