Characterization of Surface Cracks Using Eddy Current NDT Simulation by 3D-FEM and Inversion by Neural Network

摘要

In this work, we suggest an approach of signal inversion from sensors used in eddy current (EC) nondestructive testing (NDT). The aim is to characterize surface cracks from the EC signal. A methodology that combines 3D finite element (FEM) simulation and a data inversion by neural networks (NN) is proposed. We show that the use of a set of numerical measurements representing the EC signature of surface crack enables to remedy of the unicity problem. The obtained results show that the developed approach leads to the quantification of the crack.