TIME AND FREQUENCY DOMAIN FEATURE FUSION FOR DEFECT CLASSIFICATION BASED ON PULSED EDDY CURRENT NDT

摘要

Pulsed Eddy current (PEC) is an emerging nondestructive evaluation (NDT) technique excited with a broadband pulse that consists of rich frequency information. This technique generally uses the response peak value and peak arrival time to detect and quantify defects, which is not sufficient to achieve enough information associated with flaw quantification. This article introduces the application of empirical mode decomposition and Hilbert transform in extracting a new frequency domain feature from PEC responses. Time and fre-quency features are combined to enhance the accuracy of defect classification. The vali-dation and robustness of the new feature and the improved classification by experimental studies are presented.