Thickness of metallic layers from pulsed eddy-current nondestructive measurements

摘要

By using pulsed eddy-current (PEC), a neural network rnwas adopted to forecast the thickness of metal and rnthe distance rnbetween coil and metal. The relationship between the rntransient current induced and different metallic rnthickness was analyzed rntheoretically. The result shows that three features rnof the signal including the peak height, the time of rnoccurrence of the first peak, and rna zero-crossing time, are obtained. The measurement rnerror is within 2%.