Monitoring of chemical processes using improved multiscale KPCA

M. Ziyan Sheriff, M. Nazmul Karim, Mohamed Nounou, Hazem Nounou, Majdi Mansouri

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

1 Citation (Scopus)

Abstract

Statistical process monitoring charts are critical in ensuring safety for many chemical processes. Principal Component Analysis (PCA) is often used, due to its computational simplicity. However, many chemical processes may be inherently nonlinear, and this degrades the performance of the linear PCA method. Kernel Principal Component Analysis (KPCA) is an extension of the conventional PCA chart, which can help deal with nonlinearity in a given process. Additionally, PCA assumes that process data are Gaussian and uncorrelated, and only contain a moderate level of noise. These assumptions do not usually hold in practice. Multiscale wavelet-based data representation produces wavelet coefficients that possess characteristics that are able to handle violations in these assumptions. A multiscale kernel principal component analysis (MSKPCA) method has already been developed to tackle all of these issues, but it usually provides a high false alarm rate. In this paper, an improved MKSPCA chart is developed in order to deal with the false alarm rate issue, by smoothening the detection statistic using a mean filter. The advantages brought forward by the improved method are demonstrated through a simulated example in which the developed fault detection method is used to monitor a continuous stirred tank reactor (CSTR). The results clearly show that the improved MSKPCA method provides lower missed detection and false alarm rates as well as ARL1 values compared to those provided by the conventional methods.

Original languageEnglish
Title of host publication2017 4th International Conference on Control, Decision and Information Technologies, CoDIT 2017
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages49-54
Number of pages6
Volume2017-January
ISBN (Electronic)9781509064656
DOIs
Publication statusPublished - 8 Nov 2017
Event4th International Conference on Control, Decision and Information Technologies, CoDIT 2017 - Barcelona, Spain
Duration: 5 Apr 20177 Apr 2017

Other

Other4th International Conference on Control, Decision and Information Technologies, CoDIT 2017
CountrySpain
CityBarcelona
Period5/4/177/4/17

Fingerprint

Kernel Principal Component Analysis
Chemical Processes
Principal component analysis
Monitoring
Principal Component Analysis
False Alarm Rate
Chart
Process Monitoring
Wavelet Coefficients
Fault Detection
Process monitoring
Reactor
Statistic
Kernel
Fault detection
Simplicity
Monitor
Wavelets
Safety
Nonlinearity

Keywords

  • Fault detection
  • Kernel principal component analysis
  • Multiscale
  • Process monitoring
  • Wavelets

ASJC Scopus subject areas

  • Control and Optimization
  • Computational Theory and Mathematics
  • Computer Science Applications
  • Information Systems
  • Information Systems and Management
  • Artificial Intelligence

Cite this

Ziyan Sheriff, M., Nazmul Karim, M., Nounou, M., Nounou, H., & Mansouri, M. (2017). Monitoring of chemical processes using improved multiscale KPCA. In 2017 4th International Conference on Control, Decision and Information Technologies, CoDIT 2017 (Vol. 2017-January, pp. 49-54). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/CoDIT.2017.8102565

Monitoring of chemical processes using improved multiscale KPCA. / Ziyan Sheriff, M.; Nazmul Karim, M.; Nounou, Mohamed; Nounou, Hazem; Mansouri, Majdi.

2017 4th International Conference on Control, Decision and Information Technologies, CoDIT 2017. Vol. 2017-January Institute of Electrical and Electronics Engineers Inc., 2017. p. 49-54.

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

Ziyan Sheriff, M, Nazmul Karim, M, Nounou, M, Nounou, H & Mansouri, M 2017, Monitoring of chemical processes using improved multiscale KPCA. in 2017 4th International Conference on Control, Decision and Information Technologies, CoDIT 2017. vol. 2017-January, Institute of Electrical and Electronics Engineers Inc., pp. 49-54, 4th International Conference on Control, Decision and Information Technologies, CoDIT 2017, Barcelona, Spain, 5/4/17. https://doi.org/10.1109/CoDIT.2017.8102565
Ziyan Sheriff M, Nazmul Karim M, Nounou M, Nounou H, Mansouri M. Monitoring of chemical processes using improved multiscale KPCA. In 2017 4th International Conference on Control, Decision and Information Technologies, CoDIT 2017. Vol. 2017-January. Institute of Electrical and Electronics Engineers Inc. 2017. p. 49-54 https://doi.org/10.1109/CoDIT.2017.8102565
Ziyan Sheriff, M. ; Nazmul Karim, M. ; Nounou, Mohamed ; Nounou, Hazem ; Mansouri, Majdi. / Monitoring of chemical processes using improved multiscale KPCA. 2017 4th International Conference on Control, Decision and Information Technologies, CoDIT 2017. Vol. 2017-January Institute of Electrical and Electronics Engineers Inc., 2017. pp. 49-54
@inproceedings{3057fa7276ed4a259a9d5fb925c33473,
title = "Monitoring of chemical processes using improved multiscale KPCA",
abstract = "Statistical process monitoring charts are critical in ensuring safety for many chemical processes. Principal Component Analysis (PCA) is often used, due to its computational simplicity. However, many chemical processes may be inherently nonlinear, and this degrades the performance of the linear PCA method. Kernel Principal Component Analysis (KPCA) is an extension of the conventional PCA chart, which can help deal with nonlinearity in a given process. Additionally, PCA assumes that process data are Gaussian and uncorrelated, and only contain a moderate level of noise. These assumptions do not usually hold in practice. Multiscale wavelet-based data representation produces wavelet coefficients that possess characteristics that are able to handle violations in these assumptions. A multiscale kernel principal component analysis (MSKPCA) method has already been developed to tackle all of these issues, but it usually provides a high false alarm rate. In this paper, an improved MKSPCA chart is developed in order to deal with the false alarm rate issue, by smoothening the detection statistic using a mean filter. The advantages brought forward by the improved method are demonstrated through a simulated example in which the developed fault detection method is used to monitor a continuous stirred tank reactor (CSTR). The results clearly show that the improved MSKPCA method provides lower missed detection and false alarm rates as well as ARL1 values compared to those provided by the conventional methods.",
keywords = "Fault detection, Kernel principal component analysis, Multiscale, Process monitoring, Wavelets",
author = "{Ziyan Sheriff}, M. and {Nazmul Karim}, M. and Mohamed Nounou and Hazem Nounou and Majdi Mansouri",
year = "2017",
month = "11",
day = "8",
doi = "10.1109/CoDIT.2017.8102565",
language = "English",
volume = "2017-January",
pages = "49--54",
booktitle = "2017 4th International Conference on Control, Decision and Information Technologies, CoDIT 2017",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Monitoring of chemical processes using improved multiscale KPCA

AU - Ziyan Sheriff, M.

AU - Nazmul Karim, M.

AU - Nounou, Mohamed

AU - Nounou, Hazem

AU - Mansouri, Majdi

PY - 2017/11/8

Y1 - 2017/11/8

N2 - Statistical process monitoring charts are critical in ensuring safety for many chemical processes. Principal Component Analysis (PCA) is often used, due to its computational simplicity. However, many chemical processes may be inherently nonlinear, and this degrades the performance of the linear PCA method. Kernel Principal Component Analysis (KPCA) is an extension of the conventional PCA chart, which can help deal with nonlinearity in a given process. Additionally, PCA assumes that process data are Gaussian and uncorrelated, and only contain a moderate level of noise. These assumptions do not usually hold in practice. Multiscale wavelet-based data representation produces wavelet coefficients that possess characteristics that are able to handle violations in these assumptions. A multiscale kernel principal component analysis (MSKPCA) method has already been developed to tackle all of these issues, but it usually provides a high false alarm rate. In this paper, an improved MKSPCA chart is developed in order to deal with the false alarm rate issue, by smoothening the detection statistic using a mean filter. The advantages brought forward by the improved method are demonstrated through a simulated example in which the developed fault detection method is used to monitor a continuous stirred tank reactor (CSTR). The results clearly show that the improved MSKPCA method provides lower missed detection and false alarm rates as well as ARL1 values compared to those provided by the conventional methods.

AB - Statistical process monitoring charts are critical in ensuring safety for many chemical processes. Principal Component Analysis (PCA) is often used, due to its computational simplicity. However, many chemical processes may be inherently nonlinear, and this degrades the performance of the linear PCA method. Kernel Principal Component Analysis (KPCA) is an extension of the conventional PCA chart, which can help deal with nonlinearity in a given process. Additionally, PCA assumes that process data are Gaussian and uncorrelated, and only contain a moderate level of noise. These assumptions do not usually hold in practice. Multiscale wavelet-based data representation produces wavelet coefficients that possess characteristics that are able to handle violations in these assumptions. A multiscale kernel principal component analysis (MSKPCA) method has already been developed to tackle all of these issues, but it usually provides a high false alarm rate. In this paper, an improved MKSPCA chart is developed in order to deal with the false alarm rate issue, by smoothening the detection statistic using a mean filter. The advantages brought forward by the improved method are demonstrated through a simulated example in which the developed fault detection method is used to monitor a continuous stirred tank reactor (CSTR). The results clearly show that the improved MSKPCA method provides lower missed detection and false alarm rates as well as ARL1 values compared to those provided by the conventional methods.

KW - Fault detection

KW - Kernel principal component analysis

KW - Multiscale

KW - Process monitoring

KW - Wavelets

UR - http://www.scopus.com/inward/record.url?scp=85045574462&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85045574462&partnerID=8YFLogxK

U2 - 10.1109/CoDIT.2017.8102565

DO - 10.1109/CoDIT.2017.8102565

M3 - Conference contribution

VL - 2017-January

SP - 49

EP - 54

BT - 2017 4th International Conference on Control, Decision and Information Technologies, CoDIT 2017

PB - Institute of Electrical and Electronics Engineers Inc.

ER -