Evaluation of 2D Acoustic Signal Representations for Acoustic-Based Machine Condition Monitoring

Gbanaibolou Jombo; Ajay Shriram

Evaluation of 2D Acoustic Signal Representations for Acoustic-Based Machine Condition Monitoring

Research output: Contribution to conference › Paper › peer-review

33 Downloads (Pure)

Abstract

Acoustic-based machine condition monitoring (MCM) provides an improved alternative to conventional MCM approaches, including vibration analysis and lubrication monitoring, among others. Several challenges arise in anomalous machine operating sound classification, as it requires effective 2D acoustic signal representation. This paper explores this question. A baseline convolutional neural network (CNN) is implemented and trained with rolling element bearing acoustic fault data. Three representations are considered, such as log-spectrogram, short-time Fourier transform and log-Mel spectrogram. The results establish log-Mel spectrogram and log-spectrogram, as promising candidates for further exploration.

Original language	English
Pages	1-2
Number of pages	2
Publication status	Published - 12 Apr 2022
Event	PECS 2022 Physics, Engineering and Computer Science Research conference, University of Hertfordshire - University of Hertfordshire, School of Physics, Engineering and Computer Science (online), Hatfield, United Kingdom Duration: 12 Apr 2022 → 12 Apr 2022

Conference

Conference	PECS 2022 Physics, Engineering and Computer Science Research conference, University of Hertfordshire
Country/Territory	United Kingdom
City	Hatfield
Period	12/04/22 → 12/04/22

Keywords

Machine Condition Monitoring
Detection and Classification of Anomalous Machine Operating Sound
Industrial Sound Analysis
Machine Hearing

Access to Document

Final_VersionFinal published version, 389 KBLicence: CC BY

Cite this

@conference{1b1c581c0d9b4e8383314a7f7b8ab6ac,

title = "Evaluation of 2D Acoustic Signal Representations for Acoustic-Based Machine Condition Monitoring",

abstract = "Acoustic-based machine condition monitoring (MCM) provides an improved alternative to conventional MCM approaches, including vibration analysis and lubrication monitoring, among others. Several challenges arise in anomalous machine operating sound classification, as it requires effective 2D acoustic signal representation. This paper explores this question. A baseline convolutional neural network (CNN) is implemented and trained with rolling element bearing acoustic fault data. Three representations are considered, such as log-spectrogram, short-time Fourier transform and log-Mel spectrogram. The results establish log-Mel spectrogram and log-spectrogram, as promising candidates for further exploration.",

keywords = "Machine Condition Monitoring, Detection and Classification of Anomalous Machine Operating Sound, Industrial Sound Analysis, Machine Hearing",

author = "Gbanaibolou Jombo and Ajay Shriram",

year = "2022",

month = apr,

day = "12",

language = "English",

pages = "1--2",

note = "PECS 2022 Physics, Engineering and Computer Science Research conference, University of Hertfordshire ; Conference date: 12-04-2022 Through 12-04-2022",

}

TY - CONF

T1 - Evaluation of 2D Acoustic Signal Representations for Acoustic-Based Machine Condition Monitoring

AU - Jombo, Gbanaibolou

AU - Shriram, Ajay

PY - 2022/4/12

Y1 - 2022/4/12

N2 - Acoustic-based machine condition monitoring (MCM) provides an improved alternative to conventional MCM approaches, including vibration analysis and lubrication monitoring, among others. Several challenges arise in anomalous machine operating sound classification, as it requires effective 2D acoustic signal representation. This paper explores this question. A baseline convolutional neural network (CNN) is implemented and trained with rolling element bearing acoustic fault data. Three representations are considered, such as log-spectrogram, short-time Fourier transform and log-Mel spectrogram. The results establish log-Mel spectrogram and log-spectrogram, as promising candidates for further exploration.

AB - Acoustic-based machine condition monitoring (MCM) provides an improved alternative to conventional MCM approaches, including vibration analysis and lubrication monitoring, among others. Several challenges arise in anomalous machine operating sound classification, as it requires effective 2D acoustic signal representation. This paper explores this question. A baseline convolutional neural network (CNN) is implemented and trained with rolling element bearing acoustic fault data. Three representations are considered, such as log-spectrogram, short-time Fourier transform and log-Mel spectrogram. The results establish log-Mel spectrogram and log-spectrogram, as promising candidates for further exploration.

KW - Machine Condition Monitoring

KW - Detection and Classification of Anomalous Machine Operating Sound

KW - Industrial Sound Analysis

KW - Machine Hearing

M3 - Paper

SP - 1

EP - 2

T2 - PECS 2022 Physics, Engineering and Computer Science Research conference, University of Hertfordshire

Y2 - 12 April 2022 through 12 April 2022

ER -

Evaluation of 2D Acoustic Signal Representations for Acoustic-Based Machine Condition Monitoring

Abstract

Conference

Keywords

Access to Document

Fingerprint

Cite this