Anomaly sign detection by monitoring thousands of process values using a two-stage autoencoder

In a large-scale plant such as a nuclear power plant, thousands of process values are measured for the purpose of monitoring the plant performance and the health of various systems. It is difficult for plant operators to constantly monitor all of the process values. We present a data-driven method t...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Susumu NAITO, Yasunori TAGUCHI, Yuichi KATO, Kouta NAKATA, Ryota MIYAKE, Isaku NAGURA, Shinya TOMINAGA, Toshio AOKI
Formato:	article
Lenguaje:	EN
Publicado:	The Japan Society of Mechanical Engineers 2021
Materias:	plant monitoring big data anomaly sign detection machine learning deep learning autoencoder multivariate time series Mechanical engineering and machinery TJ1-1570
Acceso en línea:	https://doaj.org/article/b789e727c4134003b2d441bc15325451
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:b789e727c4134003b2d441bc15325451
record_format	dspace
spelling	oai:doaj.org-article:b789e727c4134003b2d441bc153254512021-11-29T06:09:58ZAnomaly sign detection by monitoring thousands of process values using a two-stage autoencoder2187-974510.1299/mej.20-00534https://doaj.org/article/b789e727c4134003b2d441bc153254512021-07-01T00:00:00Zhttps://www.jstage.jst.go.jp/article/mej/8/4/8_20-00534/_pdf/-char/enhttps://doaj.org/toc/2187-9745In a large-scale plant such as a nuclear power plant, thousands of process values are measured for the purpose of monitoring the plant performance and the health of various systems. It is difficult for plant operators to constantly monitor all of the process values. We present a data-driven method to comprehensively monitor a large number of process values and detect early signs of anomalies, including unknown events, with few false positives. In order to learn the complex changing internal state of a nuclear power plant and accurately predict the normal process values, we created a two-stage autoencoder composed of a time window autoencoder and a deviation autoencoder, which is a deep learning network structure corresponding to the characteristics of the process values. We assessed performances of the two-stage autoencoder with simulated process values of a nuclear power plant, a 1,100 MW boiling water reactor having 3,100 analog process values. In situations where it is difficult to predict the normal state (rapid operation mode change, transient state, and small fluctuations in the process values), the two-stage autoencoder properly predicted the normal process values and showed excellent performances with early detection and zero false positives, except for one case. The two-stage autoencoder would be an effective solution for comprehensive plant monitoring and early detection of anomaly signs.Susumu NAITOYasunori TAGUCHIYuichi KATOKouta NAKATARyota MIYAKEIsaku NAGURAShinya TOMINAGAToshio AOKIThe Japan Society of Mechanical Engineersarticleplant monitoringbig dataanomaly sign detectionmachine learningdeep learningautoencodermultivariate time seriesMechanical engineering and machineryTJ1-1570ENMechanical Engineering Journal, Vol 8, Iss 4, Pp 20-00534-20-00534 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	plant monitoring big data anomaly sign detection machine learning deep learning autoencoder multivariate time series Mechanical engineering and machinery TJ1-1570
spellingShingle	plant monitoring big data anomaly sign detection machine learning deep learning autoencoder multivariate time series Mechanical engineering and machinery TJ1-1570 Susumu NAITO Yasunori TAGUCHI Yuichi KATO Kouta NAKATA Ryota MIYAKE Isaku NAGURA Shinya TOMINAGA Toshio AOKI Anomaly sign detection by monitoring thousands of process values using a two-stage autoencoder
description	In a large-scale plant such as a nuclear power plant, thousands of process values are measured for the purpose of monitoring the plant performance and the health of various systems. It is difficult for plant operators to constantly monitor all of the process values. We present a data-driven method to comprehensively monitor a large number of process values and detect early signs of anomalies, including unknown events, with few false positives. In order to learn the complex changing internal state of a nuclear power plant and accurately predict the normal process values, we created a two-stage autoencoder composed of a time window autoencoder and a deviation autoencoder, which is a deep learning network structure corresponding to the characteristics of the process values. We assessed performances of the two-stage autoencoder with simulated process values of a nuclear power plant, a 1,100 MW boiling water reactor having 3,100 analog process values. In situations where it is difficult to predict the normal state (rapid operation mode change, transient state, and small fluctuations in the process values), the two-stage autoencoder properly predicted the normal process values and showed excellent performances with early detection and zero false positives, except for one case. The two-stage autoencoder would be an effective solution for comprehensive plant monitoring and early detection of anomaly signs.
format	article
author	Susumu NAITO Yasunori TAGUCHI Yuichi KATO Kouta NAKATA Ryota MIYAKE Isaku NAGURA Shinya TOMINAGA Toshio AOKI
author_facet	Susumu NAITO Yasunori TAGUCHI Yuichi KATO Kouta NAKATA Ryota MIYAKE Isaku NAGURA Shinya TOMINAGA Toshio AOKI
author_sort	Susumu NAITO
title	Anomaly sign detection by monitoring thousands of process values using a two-stage autoencoder
title_short	Anomaly sign detection by monitoring thousands of process values using a two-stage autoencoder
title_full	Anomaly sign detection by monitoring thousands of process values using a two-stage autoencoder
title_fullStr	Anomaly sign detection by monitoring thousands of process values using a two-stage autoencoder
title_full_unstemmed	Anomaly sign detection by monitoring thousands of process values using a two-stage autoencoder
title_sort	anomaly sign detection by monitoring thousands of process values using a two-stage autoencoder
publisher	The Japan Society of Mechanical Engineers
publishDate	2021
url	https://doaj.org/article/b789e727c4134003b2d441bc15325451
work_keys_str_mv	AT susumunaito anomalysigndetectionbymonitoringthousandsofprocessvaluesusingatwostageautoencoder AT yasunoritaguchi anomalysigndetectionbymonitoringthousandsofprocessvaluesusingatwostageautoencoder AT yuichikato anomalysigndetectionbymonitoringthousandsofprocessvaluesusingatwostageautoencoder AT koutanakata anomalysigndetectionbymonitoringthousandsofprocessvaluesusingatwostageautoencoder AT ryotamiyake anomalysigndetectionbymonitoringthousandsofprocessvaluesusingatwostageautoencoder AT isakunagura anomalysigndetectionbymonitoringthousandsofprocessvaluesusingatwostageautoencoder AT shinyatominaga anomalysigndetectionbymonitoringthousandsofprocessvaluesusingatwostageautoencoder AT toshioaoki anomalysigndetectionbymonitoringthousandsofprocessvaluesusingatwostageautoencoder
_version_	1718407609361367040

Anomaly sign detection by monitoring thousands of process values using a two-stage autoencoder

Ejemplares similares