Feature Extraction from Time-Series Data for Process Monitoring

Takeshi Fujiwara, Hirokazu Nishitani, Takeshi Fujiwara

Research output: Contribution to journalArticle

Abstract

A plant operator monitors time-series data of process variables to judge the state of process, diagnose abnormal states, and to identify failure origins. In this study, a new feature extraction method which extracts simultaneously both local features such as spikes and step changes, and the trend which characterizes global changes is provided from the viewpoint of process monitoring. In this method, the continuous function interpolated from the time-series data is represented by a series of inflection points first. Each time interval between two inflection points is called an episode. Then an approximation function of the time-series data is made iteratively by way of merging these episodes. This feature extraction method is also useful for compaction of a large number of process data.

Original languageEnglish
Pages (from-to)1103-1110
Number of pages8
JournalKagaku Kogaku Ronbunshu
Volume22
Issue number5
DOIs
Publication statusPublished - Jan 1 1996
Externally publishedYes

Fingerprint

Process monitoring
Feature extraction
Time series
Merging
Compaction

Keywords

  • Data Analysis
  • Feature Extraction
  • Process Monitoring
  • Process System
  • Process Trend

ASJC Scopus subject areas

  • Chemistry(all)
  • Chemical Engineering(all)

Cite this

Feature Extraction from Time-Series Data for Process Monitoring. / Fujiwara, Takeshi; Nishitani, Hirokazu; Fujiwara, Takeshi.

In: Kagaku Kogaku Ronbunshu, Vol. 22, No. 5, 01.01.1996, p. 1103-1110.

Research output: Contribution to journalArticle

Fujiwara, Takeshi ; Nishitani, Hirokazu ; Fujiwara, Takeshi. / Feature Extraction from Time-Series Data for Process Monitoring. In: Kagaku Kogaku Ronbunshu. 1996 ; Vol. 22, No. 5. pp. 1103-1110.
@article{440f6da49a0b4c10b69da2e769908ab3,
title = "Feature Extraction from Time-Series Data for Process Monitoring",
abstract = "A plant operator monitors time-series data of process variables to judge the state of process, diagnose abnormal states, and to identify failure origins. In this study, a new feature extraction method which extracts simultaneously both local features such as spikes and step changes, and the trend which characterizes global changes is provided from the viewpoint of process monitoring. In this method, the continuous function interpolated from the time-series data is represented by a series of inflection points first. Each time interval between two inflection points is called an episode. Then an approximation function of the time-series data is made iteratively by way of merging these episodes. This feature extraction method is also useful for compaction of a large number of process data.",
keywords = "Data Analysis, Feature Extraction, Process Monitoring, Process System, Process Trend",
author = "Takeshi Fujiwara and Hirokazu Nishitani and Takeshi Fujiwara",
year = "1996",
month = "1",
day = "1",
doi = "10.1252/kakoronbunshu.22.1103",
language = "English",
volume = "22",
pages = "1103--1110",
journal = "Kagaku Kogaku Ronbunshu",
issn = "0386-216X",
publisher = "Society of Chemical Engineers, Japan",
number = "5",

}

TY - JOUR

T1 - Feature Extraction from Time-Series Data for Process Monitoring

AU - Fujiwara, Takeshi

AU - Nishitani, Hirokazu

AU - Fujiwara, Takeshi

PY - 1996/1/1

Y1 - 1996/1/1

N2 - A plant operator monitors time-series data of process variables to judge the state of process, diagnose abnormal states, and to identify failure origins. In this study, a new feature extraction method which extracts simultaneously both local features such as spikes and step changes, and the trend which characterizes global changes is provided from the viewpoint of process monitoring. In this method, the continuous function interpolated from the time-series data is represented by a series of inflection points first. Each time interval between two inflection points is called an episode. Then an approximation function of the time-series data is made iteratively by way of merging these episodes. This feature extraction method is also useful for compaction of a large number of process data.

AB - A plant operator monitors time-series data of process variables to judge the state of process, diagnose abnormal states, and to identify failure origins. In this study, a new feature extraction method which extracts simultaneously both local features such as spikes and step changes, and the trend which characterizes global changes is provided from the viewpoint of process monitoring. In this method, the continuous function interpolated from the time-series data is represented by a series of inflection points first. Each time interval between two inflection points is called an episode. Then an approximation function of the time-series data is made iteratively by way of merging these episodes. This feature extraction method is also useful for compaction of a large number of process data.

KW - Data Analysis

KW - Feature Extraction

KW - Process Monitoring

KW - Process System

KW - Process Trend

UR - http://www.scopus.com/inward/record.url?scp=24044534275&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=24044534275&partnerID=8YFLogxK

U2 - 10.1252/kakoronbunshu.22.1103

DO - 10.1252/kakoronbunshu.22.1103

M3 - Article

VL - 22

SP - 1103

EP - 1110

JO - Kagaku Kogaku Ronbunshu

JF - Kagaku Kogaku Ronbunshu

SN - 0386-216X

IS - 5

ER -