Model-Based Reconstruction of 2-D Geometrical Features Using Eddy Current Testing

Abstract

This article presents a model-based eddy current (EC) testing method for the reconstruction of geometrical features, that may be hidden, through, or blind cavity in a large electrically conductive plate. Based on a distributed current source (DCS) modeling method, a perturbation model is formulated to convert the feature reconstruction problem into finding a possibility vector using a finite set of measured magnetic flux density (MFD) data and solved with the Tikhonov regularization method. The proposed method is numerically and experimentally validated and demonstrated by using the specimens with compound conductive/nonconductive geometrical features and comparing them with two published feature detection methods. Parametric studies are conducted to analyze the effects of feature shape, sensor configuration, and multi-coil excitation on reconstruction accuracy.