Abstract
This paper deals with the eddy current technique for nondestructive evaluation of the crack depth on a massive specimen used in aeronautical industry. A set of C-scan eddy current images is analyzed to reduce the noise and to select suitable features, which can be used to estimate the crack depth. Based on this study, a method relying on polynomial forward models of the relationship between crack depth and the maximum value of the sensor impedance is proposed. The least square and the non-negative least square techniques are applied to analyze the usability of proposed models. The error of obtained estimations is smaller than 10%, for almost used experimental data.
Highlights
The non-destructive evaluation (NDE) of metal massive structures by eddy current (EC) technique is widely considered in industry
Many other works focus on the improvement of the defect detection efficiency through the improvement of feature extraction or image processing algorithms
We proposed an approach to estimate the crack depth on a massive plate
Summary
The non-destructive evaluation (NDE) of metal massive structures by eddy current (EC) technique is widely considered in industry. To solve the estimation problem, we propose to build a multi-frequency behavioral forward model by analyzing the relationship between impedance of EC sensor and the crack depth. A first analysis of the images (treated as proposed in sections “Reduction of the influence of noise existing at the edges of images” and “Subtraction of the complex image background”) allowed us to see a clear link between the maximum value |Zmax| of the measured impedance module and the maximum depth p of the crack.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
