Enhanced performance of shewhart charts using multiscale representation

M. Ziyan Sheriff, Mohamed Nounou

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

2 Citations (Scopus)

Abstract

Monitoring charts play an essential role in statistical process control. Shewhart charts are commonly used due to their computational simplicity, and have seen many extensions that attempt to improve their performance. Most univariate charts operate under the assumption that data follow a normal distribution, are independent and contain only a moderate level of noise. Unfortunately, most practical data violate one or more of these assumptions. Wavelet-based multiscale representation of data possess characteristics that can help address these assumptions violations, and may be exploited to improve the performance of the conventional Shewhart chart. In this paper, a multiscale Shewhart chart is developed to deal with violation of these assumptions. The advantages brought forward by the developed multiscale Shewhart chart fault detection algorithm are illustrated through simulated examples. The results clearly demonstrate that the developed method is able to provide lower missed detection and comparable false alarm rates under violation of the above mentioned assumptions.

Original languageEnglish
Title of host publication2016 American Control Conference, ACC 2016
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages6923-6928
Number of pages6
Volume2016-July
ISBN (Electronic)9781467386821
DOIs
Publication statusPublished - 28 Jul 2016
Event2016 American Control Conference, ACC 2016 - Boston, United States
Duration: 6 Jul 20168 Jul 2016

Other

Other2016 American Control Conference, ACC 2016
CountryUnited States
CityBoston
Period6/7/168/7/16

Fingerprint

Statistical process control
Normal distribution
Fault detection
Control charts
Monitoring

Keywords

  • Multiscale
  • Shewhart charts
  • Wavelets

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Cite this

Sheriff, M. Z., & Nounou, M. (2016). Enhanced performance of shewhart charts using multiscale representation. In 2016 American Control Conference, ACC 2016 (Vol. 2016-July, pp. 6923-6928). [7526763] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ACC.2016.7526763

Enhanced performance of shewhart charts using multiscale representation. / Sheriff, M. Ziyan; Nounou, Mohamed.

2016 American Control Conference, ACC 2016. Vol. 2016-July Institute of Electrical and Electronics Engineers Inc., 2016. p. 6923-6928 7526763.

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

Sheriff, MZ & Nounou, M 2016, Enhanced performance of shewhart charts using multiscale representation. in 2016 American Control Conference, ACC 2016. vol. 2016-July, 7526763, Institute of Electrical and Electronics Engineers Inc., pp. 6923-6928, 2016 American Control Conference, ACC 2016, Boston, United States, 6/7/16. https://doi.org/10.1109/ACC.2016.7526763
Sheriff MZ, Nounou M. Enhanced performance of shewhart charts using multiscale representation. In 2016 American Control Conference, ACC 2016. Vol. 2016-July. Institute of Electrical and Electronics Engineers Inc. 2016. p. 6923-6928. 7526763 https://doi.org/10.1109/ACC.2016.7526763
Sheriff, M. Ziyan ; Nounou, Mohamed. / Enhanced performance of shewhart charts using multiscale representation. 2016 American Control Conference, ACC 2016. Vol. 2016-July Institute of Electrical and Electronics Engineers Inc., 2016. pp. 6923-6928
@inproceedings{f78374bd37394ceda93240b72c976f35,
title = "Enhanced performance of shewhart charts using multiscale representation",
abstract = "Monitoring charts play an essential role in statistical process control. Shewhart charts are commonly used due to their computational simplicity, and have seen many extensions that attempt to improve their performance. Most univariate charts operate under the assumption that data follow a normal distribution, are independent and contain only a moderate level of noise. Unfortunately, most practical data violate one or more of these assumptions. Wavelet-based multiscale representation of data possess characteristics that can help address these assumptions violations, and may be exploited to improve the performance of the conventional Shewhart chart. In this paper, a multiscale Shewhart chart is developed to deal with violation of these assumptions. The advantages brought forward by the developed multiscale Shewhart chart fault detection algorithm are illustrated through simulated examples. The results clearly demonstrate that the developed method is able to provide lower missed detection and comparable false alarm rates under violation of the above mentioned assumptions.",
keywords = "Multiscale, Shewhart charts, Wavelets",
author = "Sheriff, {M. Ziyan} and Mohamed Nounou",
year = "2016",
month = "7",
day = "28",
doi = "10.1109/ACC.2016.7526763",
language = "English",
volume = "2016-July",
pages = "6923--6928",
booktitle = "2016 American Control Conference, ACC 2016",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Enhanced performance of shewhart charts using multiscale representation

AU - Sheriff, M. Ziyan

AU - Nounou, Mohamed

PY - 2016/7/28

Y1 - 2016/7/28

N2 - Monitoring charts play an essential role in statistical process control. Shewhart charts are commonly used due to their computational simplicity, and have seen many extensions that attempt to improve their performance. Most univariate charts operate under the assumption that data follow a normal distribution, are independent and contain only a moderate level of noise. Unfortunately, most practical data violate one or more of these assumptions. Wavelet-based multiscale representation of data possess characteristics that can help address these assumptions violations, and may be exploited to improve the performance of the conventional Shewhart chart. In this paper, a multiscale Shewhart chart is developed to deal with violation of these assumptions. The advantages brought forward by the developed multiscale Shewhart chart fault detection algorithm are illustrated through simulated examples. The results clearly demonstrate that the developed method is able to provide lower missed detection and comparable false alarm rates under violation of the above mentioned assumptions.

AB - Monitoring charts play an essential role in statistical process control. Shewhart charts are commonly used due to their computational simplicity, and have seen many extensions that attempt to improve their performance. Most univariate charts operate under the assumption that data follow a normal distribution, are independent and contain only a moderate level of noise. Unfortunately, most practical data violate one or more of these assumptions. Wavelet-based multiscale representation of data possess characteristics that can help address these assumptions violations, and may be exploited to improve the performance of the conventional Shewhart chart. In this paper, a multiscale Shewhart chart is developed to deal with violation of these assumptions. The advantages brought forward by the developed multiscale Shewhart chart fault detection algorithm are illustrated through simulated examples. The results clearly demonstrate that the developed method is able to provide lower missed detection and comparable false alarm rates under violation of the above mentioned assumptions.

KW - Multiscale

KW - Shewhart charts

KW - Wavelets

UR - http://www.scopus.com/inward/record.url?scp=84992126319&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84992126319&partnerID=8YFLogxK

U2 - 10.1109/ACC.2016.7526763

DO - 10.1109/ACC.2016.7526763

M3 - Conference contribution

VL - 2016-July

SP - 6923

EP - 6928

BT - 2016 American Control Conference, ACC 2016

PB - Institute of Electrical and Electronics Engineers Inc.

ER -