Flaw Classification Algorithm for Heat Exchanger Tubes Using a Bobbin-Type Magnetic Camera

This paper presents an algorithm that estimates the presence, location, shape, and depth of flaws using a bobbin-type magnetic camera consisting of bobbin probes and a bobbin-type integrated giant magnetoresistance (GMR) sensor array (BIGiS). The presence of the flaws is determined by the lobe path of the Lissajous curves obtained from bobbin coil with respect to the applied frequency. The location of the flaw, i.e., whether it is an inner diameter (ID) or outer diameter (OD) flaw, can be determined from the rotational direction of the lobe with respect to the frequency change. The shape of the flaw is then determined from the area of the lobe and the BIGiS image. At this stage, multi-site damage can be determined from the BIGiS image. The effectiveness of the flaw classification algorithm was evaluated using various types of artificial flaws introduced into small-bore tube test specimens made of austenitic stainless steel.

Weighted LDA Based Dimensionality Reduction for Ultrasonic Flaw Signal Classification

Ultrasonic inspection is the most successful non-destructive testing technique for detection of flaws in engineering materials. Generally, discrete wavelet transform (DWT) is widely used to extract features for ultrasonic flaw signal. Directly taking the DWT coefficients as input vectors for training classifier, however, may result in computation complexity or even poor classification performanc. We propose a weighted linear discriminant analysis (WLDA) method to address the problem. In this study, DWT is first applied for the time-frequency analysis of ultrasonic flaw signals, and the wavelet coefficients are extracted as the initial features. After that, the proposed WLDA, which can effectively estimate the within-class and the between-class scatter through calculating similarity based weighting function, is used to reduce the dimension of original features. Finally, the features in new lower dimensional space are taken for flaw classification. Experimental results show that compared with the original and state of art linear discriminant analysis methods, WLDA is helpful for improving classification performance.