Data mining scheme for the characterization of AE signals in steel elements subject to fatigue cracking

Felipe Mejia, Navid Nemati, Antonio Nanni

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Early detection of fatigue crack-growth in steel structures is an ongoing challenge. Furthermore, characterization of the different stages of the fatigue lifecycle using NDE techniques is particularly difficult. AE systems have been shown to serve as early damage detection mechanisms in bridge structures. This technology, however, is fraught with noise problems and complex datasets that are difficult to interpret. This paper attempts to design and implement a data mining scheme that can classify raw AE datasets into discrete clusters using an improved variant of the popular k-means clustering algorithm. The datasets are then augmented with the class label found during clustering, and a series of rules are inferred using a C4.5 decision tree classification algorithm. An implementation of the data mining scheme is coded in MATLAB®, with data from PAC® AE systems as the input. In order to validate this procedure, data from a pencil lead break test with a concurrent noise source is fed into the data mining program. Classification using the decision tree is compared to manual classification of the pencil lead break hits. The resulting decision tree is then applied to a similar dataset in order to evaluate the generality of the resulting rule sets. Once validated, the data mining program is applied to data belonging to a steel fatigue crack-growth test. Results of this classification are discussed, and possible improvements to the data mining scheme are suggested.

Original languageEnglish
Title of host publicationProceedings of SPIE - The International Society for Optical Engineering
Volume8347
DOIs
StatePublished - May 15 2012
EventNondestructive Characterization for Composite Materials, Aerospace Engineering, Civil Infrastructure, and Homeland Security 2012 - San Diego, CA, United States
Duration: Mar 12 2012Mar 15 2012

Other

OtherNondestructive Characterization for Composite Materials, Aerospace Engineering, Civil Infrastructure, and Homeland Security 2012
CountryUnited States
CitySan Diego, CA
Period3/12/123/15/12

Fingerprint

data mining
Steel
Cracking
Fatigue
Data mining
Data Mining
steels
Fatigue of materials
Decision trees
Decision tree
Fatigue Crack Growth
Fatigue crack propagation
cracks
steel structures
bridges (structures)
Damage Detection
K-means Algorithm
Damage detection
K-means Clustering
Tree Algorithms

Keywords

  • Acoustic emission
  • Crack-growth
  • Data mining
  • Fatigue
  • Structural health monitoring

ASJC Scopus subject areas

  • Applied Mathematics
  • Computer Science Applications
  • Electrical and Electronic Engineering
  • Electronic, Optical and Magnetic Materials
  • Condensed Matter Physics

Cite this

Mejia, F., Nemati, N., & Nanni, A. (2012). Data mining scheme for the characterization of AE signals in steel elements subject to fatigue cracking. In Proceedings of SPIE - The International Society for Optical Engineering (Vol. 8347). [83471S] https://doi.org/10.1117/12.915383

Data mining scheme for the characterization of AE signals in steel elements subject to fatigue cracking. / Mejia, Felipe; Nemati, Navid; Nanni, Antonio.

Proceedings of SPIE - The International Society for Optical Engineering. Vol. 8347 2012. 83471S.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Mejia, F, Nemati, N & Nanni, A 2012, Data mining scheme for the characterization of AE signals in steel elements subject to fatigue cracking. in Proceedings of SPIE - The International Society for Optical Engineering. vol. 8347, 83471S, Nondestructive Characterization for Composite Materials, Aerospace Engineering, Civil Infrastructure, and Homeland Security 2012, San Diego, CA, United States, 3/12/12. https://doi.org/10.1117/12.915383
Mejia F, Nemati N, Nanni A. Data mining scheme for the characterization of AE signals in steel elements subject to fatigue cracking. In Proceedings of SPIE - The International Society for Optical Engineering. Vol. 8347. 2012. 83471S https://doi.org/10.1117/12.915383
Mejia, Felipe ; Nemati, Navid ; Nanni, Antonio. / Data mining scheme for the characterization of AE signals in steel elements subject to fatigue cracking. Proceedings of SPIE - The International Society for Optical Engineering. Vol. 8347 2012.
@inproceedings{c6d11c0a00ff4f9c9953c2a36a21fce1,
title = "Data mining scheme for the characterization of AE signals in steel elements subject to fatigue cracking",
abstract = "Early detection of fatigue crack-growth in steel structures is an ongoing challenge. Furthermore, characterization of the different stages of the fatigue lifecycle using NDE techniques is particularly difficult. AE systems have been shown to serve as early damage detection mechanisms in bridge structures. This technology, however, is fraught with noise problems and complex datasets that are difficult to interpret. This paper attempts to design and implement a data mining scheme that can classify raw AE datasets into discrete clusters using an improved variant of the popular k-means clustering algorithm. The datasets are then augmented with the class label found during clustering, and a series of rules are inferred using a C4.5 decision tree classification algorithm. An implementation of the data mining scheme is coded in MATLAB{\circledR}, with data from PAC{\circledR} AE systems as the input. In order to validate this procedure, data from a pencil lead break test with a concurrent noise source is fed into the data mining program. Classification using the decision tree is compared to manual classification of the pencil lead break hits. The resulting decision tree is then applied to a similar dataset in order to evaluate the generality of the resulting rule sets. Once validated, the data mining program is applied to data belonging to a steel fatigue crack-growth test. Results of this classification are discussed, and possible improvements to the data mining scheme are suggested.",
keywords = "Acoustic emission, Crack-growth, Data mining, Fatigue, Structural health monitoring",
author = "Felipe Mejia and Navid Nemati and Antonio Nanni",
year = "2012",
month = "5",
day = "15",
doi = "10.1117/12.915383",
language = "English",
isbn = "9780819490049",
volume = "8347",
booktitle = "Proceedings of SPIE - The International Society for Optical Engineering",

}

TY - GEN

T1 - Data mining scheme for the characterization of AE signals in steel elements subject to fatigue cracking

AU - Mejia, Felipe

AU - Nemati, Navid

AU - Nanni, Antonio

PY - 2012/5/15

Y1 - 2012/5/15

N2 - Early detection of fatigue crack-growth in steel structures is an ongoing challenge. Furthermore, characterization of the different stages of the fatigue lifecycle using NDE techniques is particularly difficult. AE systems have been shown to serve as early damage detection mechanisms in bridge structures. This technology, however, is fraught with noise problems and complex datasets that are difficult to interpret. This paper attempts to design and implement a data mining scheme that can classify raw AE datasets into discrete clusters using an improved variant of the popular k-means clustering algorithm. The datasets are then augmented with the class label found during clustering, and a series of rules are inferred using a C4.5 decision tree classification algorithm. An implementation of the data mining scheme is coded in MATLAB®, with data from PAC® AE systems as the input. In order to validate this procedure, data from a pencil lead break test with a concurrent noise source is fed into the data mining program. Classification using the decision tree is compared to manual classification of the pencil lead break hits. The resulting decision tree is then applied to a similar dataset in order to evaluate the generality of the resulting rule sets. Once validated, the data mining program is applied to data belonging to a steel fatigue crack-growth test. Results of this classification are discussed, and possible improvements to the data mining scheme are suggested.

AB - Early detection of fatigue crack-growth in steel structures is an ongoing challenge. Furthermore, characterization of the different stages of the fatigue lifecycle using NDE techniques is particularly difficult. AE systems have been shown to serve as early damage detection mechanisms in bridge structures. This technology, however, is fraught with noise problems and complex datasets that are difficult to interpret. This paper attempts to design and implement a data mining scheme that can classify raw AE datasets into discrete clusters using an improved variant of the popular k-means clustering algorithm. The datasets are then augmented with the class label found during clustering, and a series of rules are inferred using a C4.5 decision tree classification algorithm. An implementation of the data mining scheme is coded in MATLAB®, with data from PAC® AE systems as the input. In order to validate this procedure, data from a pencil lead break test with a concurrent noise source is fed into the data mining program. Classification using the decision tree is compared to manual classification of the pencil lead break hits. The resulting decision tree is then applied to a similar dataset in order to evaluate the generality of the resulting rule sets. Once validated, the data mining program is applied to data belonging to a steel fatigue crack-growth test. Results of this classification are discussed, and possible improvements to the data mining scheme are suggested.

KW - Acoustic emission

KW - Crack-growth

KW - Data mining

KW - Fatigue

KW - Structural health monitoring

UR - http://www.scopus.com/inward/record.url?scp=84860791114&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84860791114&partnerID=8YFLogxK

U2 - 10.1117/12.915383

DO - 10.1117/12.915383

M3 - Conference contribution

SN - 9780819490049

VL - 8347

BT - Proceedings of SPIE - The International Society for Optical Engineering

ER -