Application of back-propagation neural networks to defect characterization using eddy current testing

Xinwu Zhou; Ryoichi Urayama; Tetsuya Uchimoto; Toshiyuki Takagi

doi:10.3233/JAE-209394

Application of back-propagation neural networks to defect characterization using eddy current testing

Xinwu Zhou, Ryoichi Urayama, Tetsuya Uchimoto, Toshiyuki Takagi

Tohoku University

Research output: Contribution to journal › Article › peer-review

1 Citation (Scopus)

Abstract

Eddy current testing is widely used for the automatic detection of defects in conductive materials. However, this method is strongly affected by probe scanning conditions and requires signal analysis to be carried out by experienced inspectors. In this study, back-propagation neural networks were used to predict the depth and length of unknown slits by analyzing eddy current signals in the presence of noise caused by probe lift-off and tilting. The constructed neural networks were shown to predict the depth and length of defects with relative errors of 4.6% and 6.2%, respectively.

Original language	English
Pages (from-to)	817-825
Number of pages	9
Journal	International Journal of Applied Electromagnetics and Mechanics
Volume	64
Issue number	1-4
DOIs	https://doi.org/10.3233/JAE-209394
Publication status	Published - 2020

Keywords

artificial intelligence
back-propagation neural network
Eddy current testing

Access to Document

10.3233/JAE-209394

Cite this

@article{542793fefa7a4ad0a098722d8a1cac92,

title = "Application of back-propagation neural networks to defect characterization using eddy current testing",

abstract = "Eddy current testing is widely used for the automatic detection of defects in conductive materials. However, this method is strongly affected by probe scanning conditions and requires signal analysis to be carried out by experienced inspectors. In this study, back-propagation neural networks were used to predict the depth and length of unknown slits by analyzing eddy current signals in the presence of noise caused by probe lift-off and tilting. The constructed neural networks were shown to predict the depth and length of defects with relative errors of 4.6% and 6.2%, respectively.",

keywords = "artificial intelligence, back-propagation neural network, Eddy current testing",

author = "Xinwu Zhou and Ryoichi Urayama and Tetsuya Uchimoto and Toshiyuki Takagi",

note = "Funding Information: This work was supported in part by the Collaborative Research Project of the Institute of Fluid Science, Tohoku University. Publisher Copyright: {\textcopyright} 2020 - IOS Press and the authors. All rights reserved.",

year = "2020",

doi = "10.3233/JAE-209394",

language = "English",

volume = "64",

pages = "817--825",

journal = "International Journal of Applied Electromagnetics and Mechanics",

issn = "1383-5416",

publisher = "IOS Press BV",

number = "1-4",

}

TY - JOUR

T1 - Application of back-propagation neural networks to defect characterization using eddy current testing

AU - Zhou, Xinwu

AU - Urayama, Ryoichi

AU - Uchimoto, Tetsuya

AU - Takagi, Toshiyuki

N1 - Funding Information: This work was supported in part by the Collaborative Research Project of the Institute of Fluid Science, Tohoku University. Publisher Copyright: © 2020 - IOS Press and the authors. All rights reserved.

PY - 2020

Y1 - 2020

N2 - Eddy current testing is widely used for the automatic detection of defects in conductive materials. However, this method is strongly affected by probe scanning conditions and requires signal analysis to be carried out by experienced inspectors. In this study, back-propagation neural networks were used to predict the depth and length of unknown slits by analyzing eddy current signals in the presence of noise caused by probe lift-off and tilting. The constructed neural networks were shown to predict the depth and length of defects with relative errors of 4.6% and 6.2%, respectively.

AB - Eddy current testing is widely used for the automatic detection of defects in conductive materials. However, this method is strongly affected by probe scanning conditions and requires signal analysis to be carried out by experienced inspectors. In this study, back-propagation neural networks were used to predict the depth and length of unknown slits by analyzing eddy current signals in the presence of noise caused by probe lift-off and tilting. The constructed neural networks were shown to predict the depth and length of defects with relative errors of 4.6% and 6.2%, respectively.

KW - artificial intelligence

KW - back-propagation neural network

KW - Eddy current testing

UR - http://www.scopus.com/inward/record.url?scp=85097864949&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85097864949&partnerID=8YFLogxK

U2 - 10.3233/JAE-209394

DO - 10.3233/JAE-209394

M3 - Article

AN - SCOPUS:85097864949

SN - 1383-5416

VL - 64

SP - 817

EP - 825

JO - International Journal of Applied Electromagnetics and Mechanics

JF - International Journal of Applied Electromagnetics and Mechanics

IS - 1-4

ER -

Application of back-propagation neural networks to defect characterization using eddy current testing

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this