Novel adaptation of the spectral kurtosis for diagnosis of gearboxes in non-stationary conditions : Conference Paper

L. Gelman, S. Kolbe, B. Shaw, M. Vaidhianathasamy

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

Abstract

In this paper, the adaptation of Spectral Kurtosis (SK) is, for the first time in world-wide terms, proposed, demonstrated and experimentally validated. Raw data signals were collected from a single stage gearbox run at different speed and load conditions, after which Time Synchronous Averaging (TSA) was used to leave the classical residual signal once meshing harmonics were removed. Each data file is split into many individual realizations based on the time taken for the TSA to converge to stable values, after which the SK is calculated for each realization using the STFT. The effects of adapting SK technology parameters such as the SK resolution and SK threshold are evaluated, showing the effects on the consistent frequency bands identified throughout the realizations. Taking a baseline set of processing parameters, the probability of correct diagnosis was calculated using a three stage decision making technique including the k nearest neighbours and cluster analysis. Adaptation of the SK technology is then shown to dramatically improve the probability of correct diagnosis, highlighting that each speed and load case requires different SK resolution and SK threshold values to return the optimal results.

Original languageEnglish
Title of host publication13th International Conference on Condition Monitoring and Machinery Failure Prevention Technologies, CM 2016/MFPT 2016
PublisherBritish Institute of Non-Destructive Testing
Publication statusPublished - 2016
Externally publishedYes
Event13th International Conference on Condition Monitoring and Machinery Failure Prevention Technologies - Paris, France
Duration: 10 Oct 201612 Oct 2016

Conference

Conference13th International Conference on Condition Monitoring and Machinery Failure Prevention Technologies
Abbreviated titleCM & MFPT 2016
CountryFrance
CityParis
Period10/10/1612/10/16

Fingerprint

Spectral resolution
Cluster analysis
Frequency bands
Decision making
Processing

Cite this

Gelman, L., Kolbe, S., Shaw, B., & Vaidhianathasamy, M. (2016). Novel adaptation of the spectral kurtosis for diagnosis of gearboxes in non-stationary conditions : Conference Paper. In 13th International Conference on Condition Monitoring and Machinery Failure Prevention Technologies, CM 2016/MFPT 2016 British Institute of Non-Destructive Testing.
Gelman, L. ; Kolbe, S. ; Shaw, B. ; Vaidhianathasamy, M. / Novel adaptation of the spectral kurtosis for diagnosis of gearboxes in non-stationary conditions : Conference Paper. 13th International Conference on Condition Monitoring and Machinery Failure Prevention Technologies, CM 2016/MFPT 2016. British Institute of Non-Destructive Testing, 2016.
@inproceedings{c616842983384caf9fbdf65835f05778,
title = "Novel adaptation of the spectral kurtosis for diagnosis of gearboxes in non-stationary conditions : Conference Paper",
abstract = "In this paper, the adaptation of Spectral Kurtosis (SK) is, for the first time in world-wide terms, proposed, demonstrated and experimentally validated. Raw data signals were collected from a single stage gearbox run at different speed and load conditions, after which Time Synchronous Averaging (TSA) was used to leave the classical residual signal once meshing harmonics were removed. Each data file is split into many individual realizations based on the time taken for the TSA to converge to stable values, after which the SK is calculated for each realization using the STFT. The effects of adapting SK technology parameters such as the SK resolution and SK threshold are evaluated, showing the effects on the consistent frequency bands identified throughout the realizations. Taking a baseline set of processing parameters, the probability of correct diagnosis was calculated using a three stage decision making technique including the k nearest neighbours and cluster analysis. Adaptation of the SK technology is then shown to dramatically improve the probability of correct diagnosis, highlighting that each speed and load case requires different SK resolution and SK threshold values to return the optimal results.",
author = "L. Gelman and S. Kolbe and B. Shaw and M. Vaidhianathasamy",
year = "2016",
language = "English",
booktitle = "13th International Conference on Condition Monitoring and Machinery Failure Prevention Technologies, CM 2016/MFPT 2016",
publisher = "British Institute of Non-Destructive Testing",

}

Gelman, L, Kolbe, S, Shaw, B & Vaidhianathasamy, M 2016, Novel adaptation of the spectral kurtosis for diagnosis of gearboxes in non-stationary conditions : Conference Paper. in 13th International Conference on Condition Monitoring and Machinery Failure Prevention Technologies, CM 2016/MFPT 2016. British Institute of Non-Destructive Testing, 13th International Conference on Condition Monitoring and Machinery Failure Prevention Technologies, Paris, France, 10/10/16.

Novel adaptation of the spectral kurtosis for diagnosis of gearboxes in non-stationary conditions : Conference Paper. / Gelman, L.; Kolbe, S.; Shaw, B.; Vaidhianathasamy, M.

13th International Conference on Condition Monitoring and Machinery Failure Prevention Technologies, CM 2016/MFPT 2016. British Institute of Non-Destructive Testing, 2016.

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

TY - GEN

T1 - Novel adaptation of the spectral kurtosis for diagnosis of gearboxes in non-stationary conditions : Conference Paper

AU - Gelman, L.

AU - Kolbe, S.

AU - Shaw, B.

AU - Vaidhianathasamy, M.

PY - 2016

Y1 - 2016

N2 - In this paper, the adaptation of Spectral Kurtosis (SK) is, for the first time in world-wide terms, proposed, demonstrated and experimentally validated. Raw data signals were collected from a single stage gearbox run at different speed and load conditions, after which Time Synchronous Averaging (TSA) was used to leave the classical residual signal once meshing harmonics were removed. Each data file is split into many individual realizations based on the time taken for the TSA to converge to stable values, after which the SK is calculated for each realization using the STFT. The effects of adapting SK technology parameters such as the SK resolution and SK threshold are evaluated, showing the effects on the consistent frequency bands identified throughout the realizations. Taking a baseline set of processing parameters, the probability of correct diagnosis was calculated using a three stage decision making technique including the k nearest neighbours and cluster analysis. Adaptation of the SK technology is then shown to dramatically improve the probability of correct diagnosis, highlighting that each speed and load case requires different SK resolution and SK threshold values to return the optimal results.

AB - In this paper, the adaptation of Spectral Kurtosis (SK) is, for the first time in world-wide terms, proposed, demonstrated and experimentally validated. Raw data signals were collected from a single stage gearbox run at different speed and load conditions, after which Time Synchronous Averaging (TSA) was used to leave the classical residual signal once meshing harmonics were removed. Each data file is split into many individual realizations based on the time taken for the TSA to converge to stable values, after which the SK is calculated for each realization using the STFT. The effects of adapting SK technology parameters such as the SK resolution and SK threshold are evaluated, showing the effects on the consistent frequency bands identified throughout the realizations. Taking a baseline set of processing parameters, the probability of correct diagnosis was calculated using a three stage decision making technique including the k nearest neighbours and cluster analysis. Adaptation of the SK technology is then shown to dramatically improve the probability of correct diagnosis, highlighting that each speed and load case requires different SK resolution and SK threshold values to return the optimal results.

UR - http://www.scopus.com/inward/record.url?scp=85002555262&partnerID=8YFLogxK

M3 - Conference contribution

BT - 13th International Conference on Condition Monitoring and Machinery Failure Prevention Technologies, CM 2016/MFPT 2016

PB - British Institute of Non-Destructive Testing

ER -

Gelman L, Kolbe S, Shaw B, Vaidhianathasamy M. Novel adaptation of the spectral kurtosis for diagnosis of gearboxes in non-stationary conditions : Conference Paper. In 13th International Conference on Condition Monitoring and Machinery Failure Prevention Technologies, CM 2016/MFPT 2016. British Institute of Non-Destructive Testing. 2016