FORECAST OF MULTIVARIATE TIME SERIES SAMPLED FROM INDUSTRIAL MACHINERY SENSORS

Detalhes bibliográficos
Autor(a) principal: dos Santos, Heron Felipe Rosas
Data de Publicação: 2020
Outros Autores: da Silva, Leila Weitzel Coelho, Sobral, Ana Paula Barbosa
Tipo de documento: Artigo
Idioma: eng
Título da fonte: Brazilian Journal of Operations & Production Management (Online)
Texto Completo: https://bjopm.org.br/bjopm/article/view/607
Resumo: Goal: To evaluate the performance of a set of forecasting methods in the prediction of future values on a dataset of time series collected from sensors installed in an industrial gas turbine. Design/Methodology/Approach: Forecasting methods tested include the use of multivariate and univariate neural networks (FNN and LSTM), exponential smoothing and ARIMA models. Results: Results show that the use of ARIMA models to forecast on the dataset is the best default method to apply, and is the only forecasting method that consistently beats a simple naïve no-change model. Limitation of the investigation: There was a focus on evaluating neural networks. This limited resources available to evaluate other forecasting methods. There is no guarantee that it would not be possible to find neural networks capable of yielding better forecasts than the ones achieved by the best performing methods in this research. Practical implications: The broadest possible implications of the results are that the best default method to forecast industrial machinery time series is the use of ARIMA models. Additionally, neural networks are not capable of beating methods well stablished within the forecasting community, namely ARIMA models. Originality/Value: To the best of the authors’ knowledge, there is a scarce amount of published evaluations of multiple forecasting methods on data from real machines. This knowledge is useful for the understanding of the best forecasting methods available for the estimation of machine’s RUL using sensor time series.
id ABEPRO_0377a8266dde6779efa7d4c4c4c3018e
oai_identifier_str oai:ojs.bjopm.org.br:article/607
network_acronym_str ABEPRO
network_name_str Brazilian Journal of Operations & Production Management (Online)
repository_id_str
spelling FORECAST OF MULTIVARIATE TIME SERIES SAMPLED FROM INDUSTRIAL MACHINERY SENSORSprognosticstime seriesforecastingneural networksARIMAGoal: To evaluate the performance of a set of forecasting methods in the prediction of future values on a dataset of time series collected from sensors installed in an industrial gas turbine. Design/Methodology/Approach: Forecasting methods tested include the use of multivariate and univariate neural networks (FNN and LSTM), exponential smoothing and ARIMA models. Results: Results show that the use of ARIMA models to forecast on the dataset is the best default method to apply, and is the only forecasting method that consistently beats a simple naïve no-change model. Limitation of the investigation: There was a focus on evaluating neural networks. This limited resources available to evaluate other forecasting methods. There is no guarantee that it would not be possible to find neural networks capable of yielding better forecasts than the ones achieved by the best performing methods in this research. Practical implications: The broadest possible implications of the results are that the best default method to forecast industrial machinery time series is the use of ARIMA models. Additionally, neural networks are not capable of beating methods well stablished within the forecasting community, namely ARIMA models. Originality/Value: To the best of the authors’ knowledge, there is a scarce amount of published evaluations of multiple forecasting methods on data from real machines. This knowledge is useful for the understanding of the best forecasting methods available for the estimation of machine’s RUL using sensor time series.Brazilian Association for Industrial Engineering and Operations Management (ABEPRO)2020-02-06info:eu-repo/semantics/articleinfo:eu-repo/semantics/publishedVersionPeer-reviewed Articleapplication/pdfhttps://bjopm.org.br/bjopm/article/view/60710.14488/BJOPM.2020.010Brazilian Journal of Operations & Production Management; Vol. 17 No. 1 (2020): March, 2020; 1-122237-8960reponame:Brazilian Journal of Operations & Production Management (Online)instname:Associação Brasileira de Engenharia de Produção (ABEPRO)instacron:ABEPROenghttps://bjopm.org.br/bjopm/article/view/607/913Copyright (c) 2020 Heron Felipe Rosas dos Santos, Leila Weitzel Coelho da Silva, Ana Paula Barbosa Sobralinfo:eu-repo/semantics/openAccessdos Santos, Heron Felipe Rosasda Silva, Leila Weitzel CoelhoSobral, Ana Paula Barbosa2020-02-07T10:25:50Zoai:ojs.bjopm.org.br:article/607Revistahttps://bjopm.org.br/bjopmONGhttps://bjopm.org.br/bjopm/oaibjopm.journal@gmail.com2237-89601679-8171opendoar:2023-03-13T09:45:19.465280Brazilian Journal of Operations & Production Management (Online) - Associação Brasileira de Engenharia de Produção (ABEPRO)false
dc.title.none.fl_str_mv FORECAST OF MULTIVARIATE TIME SERIES SAMPLED FROM INDUSTRIAL MACHINERY SENSORS
title FORECAST OF MULTIVARIATE TIME SERIES SAMPLED FROM INDUSTRIAL MACHINERY SENSORS
spellingShingle FORECAST OF MULTIVARIATE TIME SERIES SAMPLED FROM INDUSTRIAL MACHINERY SENSORS
dos Santos, Heron Felipe Rosas
prognostics
time series
forecasting
neural networks
ARIMA
title_short FORECAST OF MULTIVARIATE TIME SERIES SAMPLED FROM INDUSTRIAL MACHINERY SENSORS
title_full FORECAST OF MULTIVARIATE TIME SERIES SAMPLED FROM INDUSTRIAL MACHINERY SENSORS
title_fullStr FORECAST OF MULTIVARIATE TIME SERIES SAMPLED FROM INDUSTRIAL MACHINERY SENSORS
title_full_unstemmed FORECAST OF MULTIVARIATE TIME SERIES SAMPLED FROM INDUSTRIAL MACHINERY SENSORS
title_sort FORECAST OF MULTIVARIATE TIME SERIES SAMPLED FROM INDUSTRIAL MACHINERY SENSORS
author dos Santos, Heron Felipe Rosas
author_facet dos Santos, Heron Felipe Rosas
da Silva, Leila Weitzel Coelho
Sobral, Ana Paula Barbosa
author_role author
author2 da Silva, Leila Weitzel Coelho
Sobral, Ana Paula Barbosa
author2_role author
author
dc.contributor.author.fl_str_mv dos Santos, Heron Felipe Rosas
da Silva, Leila Weitzel Coelho
Sobral, Ana Paula Barbosa
dc.subject.por.fl_str_mv prognostics
time series
forecasting
neural networks
ARIMA
topic prognostics
time series
forecasting
neural networks
ARIMA
description Goal: To evaluate the performance of a set of forecasting methods in the prediction of future values on a dataset of time series collected from sensors installed in an industrial gas turbine. Design/Methodology/Approach: Forecasting methods tested include the use of multivariate and univariate neural networks (FNN and LSTM), exponential smoothing and ARIMA models. Results: Results show that the use of ARIMA models to forecast on the dataset is the best default method to apply, and is the only forecasting method that consistently beats a simple naïve no-change model. Limitation of the investigation: There was a focus on evaluating neural networks. This limited resources available to evaluate other forecasting methods. There is no guarantee that it would not be possible to find neural networks capable of yielding better forecasts than the ones achieved by the best performing methods in this research. Practical implications: The broadest possible implications of the results are that the best default method to forecast industrial machinery time series is the use of ARIMA models. Additionally, neural networks are not capable of beating methods well stablished within the forecasting community, namely ARIMA models. Originality/Value: To the best of the authors’ knowledge, there is a scarce amount of published evaluations of multiple forecasting methods on data from real machines. This knowledge is useful for the understanding of the best forecasting methods available for the estimation of machine’s RUL using sensor time series.
publishDate 2020
dc.date.none.fl_str_mv 2020-02-06
dc.type.driver.fl_str_mv info:eu-repo/semantics/article
info:eu-repo/semantics/publishedVersion
Peer-reviewed Article
format article
status_str publishedVersion
dc.identifier.uri.fl_str_mv https://bjopm.org.br/bjopm/article/view/607
10.14488/BJOPM.2020.010
url https://bjopm.org.br/bjopm/article/view/607
identifier_str_mv 10.14488/BJOPM.2020.010
dc.language.iso.fl_str_mv eng
language eng
dc.relation.none.fl_str_mv https://bjopm.org.br/bjopm/article/view/607/913
dc.rights.driver.fl_str_mv info:eu-repo/semantics/openAccess
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv application/pdf
dc.publisher.none.fl_str_mv Brazilian Association for Industrial Engineering and Operations Management (ABEPRO)
publisher.none.fl_str_mv Brazilian Association for Industrial Engineering and Operations Management (ABEPRO)
dc.source.none.fl_str_mv Brazilian Journal of Operations & Production Management; Vol. 17 No. 1 (2020): March, 2020; 1-12
2237-8960
reponame:Brazilian Journal of Operations & Production Management (Online)
instname:Associação Brasileira de Engenharia de Produção (ABEPRO)
instacron:ABEPRO
instname_str Associação Brasileira de Engenharia de Produção (ABEPRO)
instacron_str ABEPRO
institution ABEPRO
reponame_str Brazilian Journal of Operations & Production Management (Online)
collection Brazilian Journal of Operations & Production Management (Online)
repository.name.fl_str_mv Brazilian Journal of Operations & Production Management (Online) - Associação Brasileira de Engenharia de Produção (ABEPRO)
repository.mail.fl_str_mv bjopm.journal@gmail.com
_version_ 1797051461028282368