Multivariate statistical process control-based hypothesis testing for damage detection in structural health monitoring systems

Marwa Chaabane, Majdi Mansouri, Ahmed Ben Hamida, Hazem Nounou, Mohamed Nounou

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

The objective of this paper is to propose a new damage detection technique based on multiscale partial least squares (MSPLS) and optimized exponentially weighted moving average (OEWMA) generalized likelihood ratio test (GLRT) to enhance monitoring of structural systems. The developed technique attempts to combine the advantages of the exponentially weighted moving average (EWMA) and GLRT charts with those of multiscale input-output model partial least square (PLS) and multi-objective optimization. The damage detection problem is addressed so that the data are first modeled using the MSPLS method and then the damages are detected using the OEWMA-GLRT chart. The idea behind the developed OEWMA-GLRT is to compute an optimal statistic that integrates current and previous data information in a decreasing exponential fashion giving more weight to the more recent data and selects the EWMA parameters that minimizes the (MDR), the false alarm rate (FAR) and the average run length (ARL1). This helps provide a more accurate estimation of the GLRT statistic and provide a stronger memory that enables better decision making with respect to damage detection. The performance of the developed technique is assessed and compared with PLS-based GLRT, PLS-based OEWMA, and PLS-based OEWMA-GLRT techniques using two illustrative examples, synthetic data and simulated International Association for Structural Control-American society of Civil engineers (IASC-ASCE) benchmark structure. The results demonstrate the effectiveness of the MSPLS-based OEWMA-GLRT technique over the PLS-based GLRT, PLS-based OEWMA, and PLS-based OEWMA-GLRT methods in terms of MDR, FAR, and ARL1 values.

Original languageEnglish
Article numbere2287
JournalStructural Control and Health Monitoring
DOIs
Publication statusAccepted/In press - 1 Jan 2018

Fingerprint

Statistical process control
Damage detection
Structural health monitoring
Testing
Statistics
Multiobjective optimization
Decision making
Data storage equipment
Engineers
Monitoring

Keywords

  • damage detection
  • exponentially weighted moving average
  • generalized likelihood ratio test
  • multiscale
  • partial least squares
  • structural health monitoring

ASJC Scopus subject areas

  • Civil and Structural Engineering
  • Building and Construction
  • Mechanics of Materials

Cite this

@article{a5289d460b81449ebb6f7b82e3125d15,
title = "Multivariate statistical process control-based hypothesis testing for damage detection in structural health monitoring systems",
abstract = "The objective of this paper is to propose a new damage detection technique based on multiscale partial least squares (MSPLS) and optimized exponentially weighted moving average (OEWMA) generalized likelihood ratio test (GLRT) to enhance monitoring of structural systems. The developed technique attempts to combine the advantages of the exponentially weighted moving average (EWMA) and GLRT charts with those of multiscale input-output model partial least square (PLS) and multi-objective optimization. The damage detection problem is addressed so that the data are first modeled using the MSPLS method and then the damages are detected using the OEWMA-GLRT chart. The idea behind the developed OEWMA-GLRT is to compute an optimal statistic that integrates current and previous data information in a decreasing exponential fashion giving more weight to the more recent data and selects the EWMA parameters that minimizes the (MDR), the false alarm rate (FAR) and the average run length (ARL1). This helps provide a more accurate estimation of the GLRT statistic and provide a stronger memory that enables better decision making with respect to damage detection. The performance of the developed technique is assessed and compared with PLS-based GLRT, PLS-based OEWMA, and PLS-based OEWMA-GLRT techniques using two illustrative examples, synthetic data and simulated International Association for Structural Control-American society of Civil engineers (IASC-ASCE) benchmark structure. The results demonstrate the effectiveness of the MSPLS-based OEWMA-GLRT technique over the PLS-based GLRT, PLS-based OEWMA, and PLS-based OEWMA-GLRT methods in terms of MDR, FAR, and ARL1 values.",
keywords = "damage detection, exponentially weighted moving average, generalized likelihood ratio test, multiscale, partial least squares, structural health monitoring",
author = "Marwa Chaabane and Majdi Mansouri and {Ben Hamida}, Ahmed and Hazem Nounou and Mohamed Nounou",
year = "2018",
month = "1",
day = "1",
doi = "10.1002/stc.2287",
language = "English",
journal = "Structural Control and Health Monitoring",
issn = "1545-2255",
publisher = "John Wiley and Sons Ltd",

}

TY - JOUR

T1 - Multivariate statistical process control-based hypothesis testing for damage detection in structural health monitoring systems

AU - Chaabane, Marwa

AU - Mansouri, Majdi

AU - Ben Hamida, Ahmed

AU - Nounou, Hazem

AU - Nounou, Mohamed

PY - 2018/1/1

Y1 - 2018/1/1

N2 - The objective of this paper is to propose a new damage detection technique based on multiscale partial least squares (MSPLS) and optimized exponentially weighted moving average (OEWMA) generalized likelihood ratio test (GLRT) to enhance monitoring of structural systems. The developed technique attempts to combine the advantages of the exponentially weighted moving average (EWMA) and GLRT charts with those of multiscale input-output model partial least square (PLS) and multi-objective optimization. The damage detection problem is addressed so that the data are first modeled using the MSPLS method and then the damages are detected using the OEWMA-GLRT chart. The idea behind the developed OEWMA-GLRT is to compute an optimal statistic that integrates current and previous data information in a decreasing exponential fashion giving more weight to the more recent data and selects the EWMA parameters that minimizes the (MDR), the false alarm rate (FAR) and the average run length (ARL1). This helps provide a more accurate estimation of the GLRT statistic and provide a stronger memory that enables better decision making with respect to damage detection. The performance of the developed technique is assessed and compared with PLS-based GLRT, PLS-based OEWMA, and PLS-based OEWMA-GLRT techniques using two illustrative examples, synthetic data and simulated International Association for Structural Control-American society of Civil engineers (IASC-ASCE) benchmark structure. The results demonstrate the effectiveness of the MSPLS-based OEWMA-GLRT technique over the PLS-based GLRT, PLS-based OEWMA, and PLS-based OEWMA-GLRT methods in terms of MDR, FAR, and ARL1 values.

AB - The objective of this paper is to propose a new damage detection technique based on multiscale partial least squares (MSPLS) and optimized exponentially weighted moving average (OEWMA) generalized likelihood ratio test (GLRT) to enhance monitoring of structural systems. The developed technique attempts to combine the advantages of the exponentially weighted moving average (EWMA) and GLRT charts with those of multiscale input-output model partial least square (PLS) and multi-objective optimization. The damage detection problem is addressed so that the data are first modeled using the MSPLS method and then the damages are detected using the OEWMA-GLRT chart. The idea behind the developed OEWMA-GLRT is to compute an optimal statistic that integrates current and previous data information in a decreasing exponential fashion giving more weight to the more recent data and selects the EWMA parameters that minimizes the (MDR), the false alarm rate (FAR) and the average run length (ARL1). This helps provide a more accurate estimation of the GLRT statistic and provide a stronger memory that enables better decision making with respect to damage detection. The performance of the developed technique is assessed and compared with PLS-based GLRT, PLS-based OEWMA, and PLS-based OEWMA-GLRT techniques using two illustrative examples, synthetic data and simulated International Association for Structural Control-American society of Civil engineers (IASC-ASCE) benchmark structure. The results demonstrate the effectiveness of the MSPLS-based OEWMA-GLRT technique over the PLS-based GLRT, PLS-based OEWMA, and PLS-based OEWMA-GLRT methods in terms of MDR, FAR, and ARL1 values.

KW - damage detection

KW - exponentially weighted moving average

KW - generalized likelihood ratio test

KW - multiscale

KW - partial least squares

KW - structural health monitoring

UR - http://www.scopus.com/inward/record.url?scp=85055730785&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85055730785&partnerID=8YFLogxK

U2 - 10.1002/stc.2287

DO - 10.1002/stc.2287

M3 - Article

JO - Structural Control and Health Monitoring

JF - Structural Control and Health Monitoring

SN - 1545-2255

M1 - e2287

ER -