Damage location in an isotropic plate using a vector of novelty indices

F. Mustapha, G. Manson, K. Worden, S.G. Pierce

Research output: Contribution to journalArticle

23 Citations (Scopus)

Abstract

This paper presents a novel approach to detecting and localising structural defects based on a novelty detection method, outlier analysis (OA), and a multi-layer perceptron (MLP) neural network. In order to assess the effectiveness of the approach, a thin rectangular plate with isotropic behaviour was evaluated experimentally. The scope of this present work also comprises an investigation of the scattering effect of an ultrasonic guided wave on the tested plate under both damaged and undamaged conditions. The wave propagation is sequentially transmitted and captured by 8 PZT patches bonded on the plate, forming a sensor network on the tested isotropic rectangular structure. An in-house 8 channel multiplexer is incorporated in this small scale and low-cost ‘structural health monitoring’ (SHM) system to effectively swap the PZTs role from sensor to actuator and vice-versa. The ‘real-time damage demonstrator’ software is primarily developed to acquire and store the waveform responses. These sets of scattering waveform responses representing normal and damage conditions are transformed into a set of novelty indices that ultimately determine the true conditions of the tested structure. The acquired novelty indices representing the available sensor paths are used as the inputs for the neural network incorporating the MLP architecture to compute and predict the damage location in the x and y location on the tested isotopic plate.
LanguageEnglish
Pages1885-1906
Number of pages22
JournalMechanical Systems and Signal Processing
Volume21
Issue number4
DOIs
Publication statusPublished - May 2007

Fingerprint

Multilayer neural networks
Scattering
Neural networks
Guided electromagnetic wave propagation
Structural health monitoring
Ultrasonic waves
Sensors
Wave propagation
Sensor networks
Actuators
Defects
Costs

Keywords

  • structural health monitoring
  • outlier analysis
  • multi-layer perceptron
  • neural network
  • ultrasonic guided wave
  • isotropic plate

Cite this

Mustapha, F. ; Manson, G. ; Worden, K. ; Pierce, S.G. / Damage location in an isotropic plate using a vector of novelty indices. In: Mechanical Systems and Signal Processing. 2007 ; Vol. 21, No. 4. pp. 1885-1906.
@article{6e052487ba4846b3afc663410e7a2e56,
title = "Damage location in an isotropic plate using a vector of novelty indices",
abstract = "This paper presents a novel approach to detecting and localising structural defects based on a novelty detection method, outlier analysis (OA), and a multi-layer perceptron (MLP) neural network. In order to assess the effectiveness of the approach, a thin rectangular plate with isotropic behaviour was evaluated experimentally. The scope of this present work also comprises an investigation of the scattering effect of an ultrasonic guided wave on the tested plate under both damaged and undamaged conditions. The wave propagation is sequentially transmitted and captured by 8 PZT patches bonded on the plate, forming a sensor network on the tested isotropic rectangular structure. An in-house 8 channel multiplexer is incorporated in this small scale and low-cost ‘structural health monitoring’ (SHM) system to effectively swap the PZTs role from sensor to actuator and vice-versa. The ‘real-time damage demonstrator’ software is primarily developed to acquire and store the waveform responses. These sets of scattering waveform responses representing normal and damage conditions are transformed into a set of novelty indices that ultimately determine the true conditions of the tested structure. The acquired novelty indices representing the available sensor paths are used as the inputs for the neural network incorporating the MLP architecture to compute and predict the damage location in the x and y location on the tested isotopic plate.",
keywords = "structural health monitoring, outlier analysis, multi-layer perceptron, neural network, ultrasonic guided wave, isotropic plate",
author = "F. Mustapha and G. Manson and K. Worden and S.G. Pierce",
year = "2007",
month = "5",
doi = "10.1016/j.ymssp.2006.06.009",
language = "English",
volume = "21",
pages = "1885--1906",
journal = "Mechanical Systems and Signal Processing",
issn = "0888-3270",
number = "4",

}

Damage location in an isotropic plate using a vector of novelty indices. / Mustapha, F.; Manson, G.; Worden, K.; Pierce, S.G.

In: Mechanical Systems and Signal Processing, Vol. 21, No. 4, 05.2007, p. 1885-1906.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Damage location in an isotropic plate using a vector of novelty indices

AU - Mustapha, F.

AU - Manson, G.

AU - Worden, K.

AU - Pierce, S.G.

PY - 2007/5

Y1 - 2007/5

N2 - This paper presents a novel approach to detecting and localising structural defects based on a novelty detection method, outlier analysis (OA), and a multi-layer perceptron (MLP) neural network. In order to assess the effectiveness of the approach, a thin rectangular plate with isotropic behaviour was evaluated experimentally. The scope of this present work also comprises an investigation of the scattering effect of an ultrasonic guided wave on the tested plate under both damaged and undamaged conditions. The wave propagation is sequentially transmitted and captured by 8 PZT patches bonded on the plate, forming a sensor network on the tested isotropic rectangular structure. An in-house 8 channel multiplexer is incorporated in this small scale and low-cost ‘structural health monitoring’ (SHM) system to effectively swap the PZTs role from sensor to actuator and vice-versa. The ‘real-time damage demonstrator’ software is primarily developed to acquire and store the waveform responses. These sets of scattering waveform responses representing normal and damage conditions are transformed into a set of novelty indices that ultimately determine the true conditions of the tested structure. The acquired novelty indices representing the available sensor paths are used as the inputs for the neural network incorporating the MLP architecture to compute and predict the damage location in the x and y location on the tested isotopic plate.

AB - This paper presents a novel approach to detecting and localising structural defects based on a novelty detection method, outlier analysis (OA), and a multi-layer perceptron (MLP) neural network. In order to assess the effectiveness of the approach, a thin rectangular plate with isotropic behaviour was evaluated experimentally. The scope of this present work also comprises an investigation of the scattering effect of an ultrasonic guided wave on the tested plate under both damaged and undamaged conditions. The wave propagation is sequentially transmitted and captured by 8 PZT patches bonded on the plate, forming a sensor network on the tested isotropic rectangular structure. An in-house 8 channel multiplexer is incorporated in this small scale and low-cost ‘structural health monitoring’ (SHM) system to effectively swap the PZTs role from sensor to actuator and vice-versa. The ‘real-time damage demonstrator’ software is primarily developed to acquire and store the waveform responses. These sets of scattering waveform responses representing normal and damage conditions are transformed into a set of novelty indices that ultimately determine the true conditions of the tested structure. The acquired novelty indices representing the available sensor paths are used as the inputs for the neural network incorporating the MLP architecture to compute and predict the damage location in the x and y location on the tested isotopic plate.

KW - structural health monitoring

KW - outlier analysis

KW - multi-layer perceptron

KW - neural network

KW - ultrasonic guided wave

KW - isotropic plate

U2 - 10.1016/j.ymssp.2006.06.009

DO - 10.1016/j.ymssp.2006.06.009

M3 - Article

VL - 21

SP - 1885

EP - 1906

JO - Mechanical Systems and Signal Processing

T2 - Mechanical Systems and Signal Processing

JF - Mechanical Systems and Signal Processing

SN - 0888-3270

IS - 4

ER -