The real-time retrieval of the characteristics of a defect with eddy current testing in a nondestructive testing and evaluation framework is addressed. An innovative statistical learning approach is developed to deal with the inversion problem at hand in a computationally efficient way. More in detail, a feature extraction technique based on partial least squares (PLS) is profitably combined with a customized output space filling (OSF) adaptive sampling scheme for generating optimal training databases, while accurate and robust reconstructions are performed with a support vector regression (SVR) algorithm. A selected set of numerical and experimental results is reported to assess the effectiveness as well as the efficiency of the proposed PLS-OSF/SVR approach.

Real-Time NDT-NDE Through an Innovative Adaptive Partial Least Squares SVR Inversion Approach

Salucci, Marco;Anselmi, Nicola;Oliveri, Giacomo;Massa, Andrea
2016-01-01

Abstract

The real-time retrieval of the characteristics of a defect with eddy current testing in a nondestructive testing and evaluation framework is addressed. An innovative statistical learning approach is developed to deal with the inversion problem at hand in a computationally efficient way. More in detail, a feature extraction technique based on partial least squares (PLS) is profitably combined with a customized output space filling (OSF) adaptive sampling scheme for generating optimal training databases, while accurate and robust reconstructions are performed with a support vector regression (SVR) algorithm. A selected set of numerical and experimental results is reported to assess the effectiveness as well as the efficiency of the proposed PLS-OSF/SVR approach.
2016
11
Salucci, Marco; Anselmi, Nicola; Oliveri, Giacomo; Calmon, Pierre; Miorelli, Roberto; Reboud, Christophe; Massa, Andrea
File in questo prodotto:
File Dimensione Formato  
R282.pdf

Solo gestori archivio

Tipologia: Versione editoriale (Publisher’s layout)
Licenza: Tutti i diritti riservati (All rights reserved)
Dimensione 6.36 MB
Formato Adobe PDF
6.36 MB Adobe PDF   Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11572/155101
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 76
  • ???jsp.display-item.citation.isi??? 62
social impact