The aim of this paper is to present a neural network based solution to a non destructive test problem, namely the identification of the diameter of a cylindrical defect, on a metallic slab, by means of a multilayer perceptron based model of the complex interaction between the metallic slab and the electromagnetic probe. We propose to train a network by means of a consistent data set obtained by a real-world (measured) data, labeled with the defect diameters, and to successivelly apply the learn network to the estimation of the dimention of a set of unknown defects.
A Multilayer Percepron Approach to a Non-Destructive Test Problem
FABA, Antonio;BURRASCANO, Pietro;CARDELLI, Ermanno;
2001
Abstract
The aim of this paper is to present a neural network based solution to a non destructive test problem, namely the identification of the diameter of a cylindrical defect, on a metallic slab, by means of a multilayer perceptron based model of the complex interaction between the metallic slab and the electromagnetic probe. We propose to train a network by means of a consistent data set obtained by a real-world (measured) data, labeled with the defect diameters, and to successivelly apply the learn network to the estimation of the dimention of a set of unknown defects.File in questo prodotto:
Non ci sono file associati a questo prodotto.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
