Crack Identification by Artificial Neural Network

Chyanbin Hwu, Y. C. Liang

Research output: Contribution to journalArticle

6 Citations (Scopus)

Abstract

In this paper, a most popular artificial neural network called the back propagation neural network (BPN) is employed to achieve an ideal on-line identification of the crack embedded in a composite plate. Different from the usual dynamic estimate, the parameters used for the present crack identification are the strains of static deformation. It is known that the crack effects are localized which may not be clearly reflected from the boundary information especially when the data is from static deformation only. To remedy this, we use data from multiple-loading modes in which the loading modes may include the opening, shearing and tearing modes. The results show that our method for crack identification is always stable and accurate no matter how far-away of the test data from its training set.

Original languageEnglish
Pages (from-to)405-410
Number of pages6
JournalKey Engineering Materials
Volume145-149
Publication statusPublished - 1998 Dec 1

Fingerprint

Cracks
Neural networks
Backpropagation
Shearing
Composite materials

All Science Journal Classification (ASJC) codes

  • Materials Science(all)
  • Mechanics of Materials
  • Mechanical Engineering

Cite this

Hwu, Chyanbin ; Liang, Y. C. / Crack Identification by Artificial Neural Network. In: Key Engineering Materials. 1998 ; Vol. 145-149. pp. 405-410.
@article{a813edfbc5f94e31a64ad1e0a19dbf81,
title = "Crack Identification by Artificial Neural Network",
abstract = "In this paper, a most popular artificial neural network called the back propagation neural network (BPN) is employed to achieve an ideal on-line identification of the crack embedded in a composite plate. Different from the usual dynamic estimate, the parameters used for the present crack identification are the strains of static deformation. It is known that the crack effects are localized which may not be clearly reflected from the boundary information especially when the data is from static deformation only. To remedy this, we use data from multiple-loading modes in which the loading modes may include the opening, shearing and tearing modes. The results show that our method for crack identification is always stable and accurate no matter how far-away of the test data from its training set.",
author = "Chyanbin Hwu and Liang, {Y. C.}",
year = "1998",
month = "12",
day = "1",
language = "English",
volume = "145-149",
pages = "405--410",
journal = "Key Engineering Materials",
issn = "1013-9826",
publisher = "Trans Tech Publications",

}

Crack Identification by Artificial Neural Network. / Hwu, Chyanbin; Liang, Y. C.

In: Key Engineering Materials, Vol. 145-149, 01.12.1998, p. 405-410.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Crack Identification by Artificial Neural Network

AU - Hwu, Chyanbin

AU - Liang, Y. C.

PY - 1998/12/1

Y1 - 1998/12/1

N2 - In this paper, a most popular artificial neural network called the back propagation neural network (BPN) is employed to achieve an ideal on-line identification of the crack embedded in a composite plate. Different from the usual dynamic estimate, the parameters used for the present crack identification are the strains of static deformation. It is known that the crack effects are localized which may not be clearly reflected from the boundary information especially when the data is from static deformation only. To remedy this, we use data from multiple-loading modes in which the loading modes may include the opening, shearing and tearing modes. The results show that our method for crack identification is always stable and accurate no matter how far-away of the test data from its training set.

AB - In this paper, a most popular artificial neural network called the back propagation neural network (BPN) is employed to achieve an ideal on-line identification of the crack embedded in a composite plate. Different from the usual dynamic estimate, the parameters used for the present crack identification are the strains of static deformation. It is known that the crack effects are localized which may not be clearly reflected from the boundary information especially when the data is from static deformation only. To remedy this, we use data from multiple-loading modes in which the loading modes may include the opening, shearing and tearing modes. The results show that our method for crack identification is always stable and accurate no matter how far-away of the test data from its training set.

UR - http://www.scopus.com/inward/record.url?scp=0003716857&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0003716857&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:0003716857

VL - 145-149

SP - 405

EP - 410

JO - Key Engineering Materials

JF - Key Engineering Materials

SN - 1013-9826

ER -