Two-stage damage diagnosis based on the distance between ARMA models and pre-whitening filters

H. Zheng, Akira Mita

Research output: Contribution to journalArticle

39 Citations (Scopus)

Abstract

This paper presents a two-stage damage diagnosis strategy for damage detection and localization. Auto-regressive moving-average (ARMA) models are fitted to time series of vibration signals recorded by sensors. In the first stage, a novel damage indicator, which is defined as the distance between ARMA models, is applied to damage detection. This stage can determine the existence of damage in the structure. Such an algorithm uses output only and does not require operator intervention. Therefore it can be embedded in the sensor board of a monitoring network. In the second stage, a pre-whitening filter is used to minimize the cross-correlation of multiple excitations. With this technique, the damage indicator can further identify the damage location and severity when the damage has been detected in the first stage. The proposed methodology is tested using simulation and experimental data. The analysis results clearly illustrate the feasibility of the proposed two-stage damage diagnosis methodology.

Original languageEnglish
Pages (from-to)1829-1836
Number of pages8
JournalSmart Materials and Structures
Volume16
Issue number5
DOIs
Publication statusPublished - 2007 Oct 1

Fingerprint

autoregressive moving average
Damage detection
damage
filters
Sensors
Mathematical operators
Time series
Monitoring
methodology
sensors
cross correlation
operators
vibration

ASJC Scopus subject areas

  • Materials Science(all)

Cite this

Two-stage damage diagnosis based on the distance between ARMA models and pre-whitening filters. / Zheng, H.; Mita, Akira.

In: Smart Materials and Structures, Vol. 16, No. 5, 01.10.2007, p. 1829-1836.

Research output: Contribution to journalArticle

@article{42b47170eb594f39bbca208659ab1a7d,
title = "Two-stage damage diagnosis based on the distance between ARMA models and pre-whitening filters",
abstract = "This paper presents a two-stage damage diagnosis strategy for damage detection and localization. Auto-regressive moving-average (ARMA) models are fitted to time series of vibration signals recorded by sensors. In the first stage, a novel damage indicator, which is defined as the distance between ARMA models, is applied to damage detection. This stage can determine the existence of damage in the structure. Such an algorithm uses output only and does not require operator intervention. Therefore it can be embedded in the sensor board of a monitoring network. In the second stage, a pre-whitening filter is used to minimize the cross-correlation of multiple excitations. With this technique, the damage indicator can further identify the damage location and severity when the damage has been detected in the first stage. The proposed methodology is tested using simulation and experimental data. The analysis results clearly illustrate the feasibility of the proposed two-stage damage diagnosis methodology.",
author = "H. Zheng and Akira Mita",
year = "2007",
month = "10",
day = "1",
doi = "10.1088/0964-1726/16/5/038",
language = "English",
volume = "16",
pages = "1829--1836",
journal = "Smart Materials and Structures",
issn = "0964-1726",
publisher = "IOP Publishing Ltd.",
number = "5",

}

TY - JOUR

T1 - Two-stage damage diagnosis based on the distance between ARMA models and pre-whitening filters

AU - Zheng, H.

AU - Mita, Akira

PY - 2007/10/1

Y1 - 2007/10/1

N2 - This paper presents a two-stage damage diagnosis strategy for damage detection and localization. Auto-regressive moving-average (ARMA) models are fitted to time series of vibration signals recorded by sensors. In the first stage, a novel damage indicator, which is defined as the distance between ARMA models, is applied to damage detection. This stage can determine the existence of damage in the structure. Such an algorithm uses output only and does not require operator intervention. Therefore it can be embedded in the sensor board of a monitoring network. In the second stage, a pre-whitening filter is used to minimize the cross-correlation of multiple excitations. With this technique, the damage indicator can further identify the damage location and severity when the damage has been detected in the first stage. The proposed methodology is tested using simulation and experimental data. The analysis results clearly illustrate the feasibility of the proposed two-stage damage diagnosis methodology.

AB - This paper presents a two-stage damage diagnosis strategy for damage detection and localization. Auto-regressive moving-average (ARMA) models are fitted to time series of vibration signals recorded by sensors. In the first stage, a novel damage indicator, which is defined as the distance between ARMA models, is applied to damage detection. This stage can determine the existence of damage in the structure. Such an algorithm uses output only and does not require operator intervention. Therefore it can be embedded in the sensor board of a monitoring network. In the second stage, a pre-whitening filter is used to minimize the cross-correlation of multiple excitations. With this technique, the damage indicator can further identify the damage location and severity when the damage has been detected in the first stage. The proposed methodology is tested using simulation and experimental data. The analysis results clearly illustrate the feasibility of the proposed two-stage damage diagnosis methodology.

UR - http://www.scopus.com/inward/record.url?scp=39749154165&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=39749154165&partnerID=8YFLogxK

U2 - 10.1088/0964-1726/16/5/038

DO - 10.1088/0964-1726/16/5/038

M3 - Article

VL - 16

SP - 1829

EP - 1836

JO - Smart Materials and Structures

JF - Smart Materials and Structures

SN - 0964-1726

IS - 5

ER -