Defect classification based on rectangular pulsed eddy current sensor in different directions

Abstract

Pulsed eddy current (PEC) testing is a new emerging and effective electromagnetic non-destructive testing (NDT) technique. The main purpose of this study is to identify surface defects and sub-surface defects using features-based rectangular pulsed eddy current sensor. The further study of PEC rectangular sensor proposed in author's previous work has been made to classify the different types of defects in specimen. In different directions of sensor scanning, peak waves of pick-up coil are studied. We find that when sensor is on different position against the defect, peak waves of response signals present the same shape in direction of magnetic induction flux, while present different shapes in direction of exciting current. Experiment results have shown that the different classes of defects can be identified and classified effectively by selecting the rising time as the time domain feature in both directions. For improving the performance of defect classification, two new features from differential response signal are proposed to classify different types of defects combined with rising time. One is called as crossing time; the other is differential time to peak. The blind test is carried out and the results show that the new features are effective to classify the defects.