Dynamical Analysis for Flaw Shape Identification in non Linear Eddy Current Tests

Purpose – The paper seeks to propose a specific approach based on Dynamic Analysis and Chaos Theory aiming to emphasize the differences into the eddy current signals obtained by related non-destructive tests, when the inspected specimens have flaws with different shapes. Design/methodology/approach – Non-linear eddy current analysis is very useful for flaw detection in many in-service inspections. State-of-the-art technologies allow one to define position and depth of defects, but the shape identification is still an open problem. In this paper, experimental data have been subjected to a dynamical analysis in order to relate the trend of eddy current signals to the shape of analyzed defect. Findings – In particular, a dynamical reconstruction by means of recurrence plots (RPs) has been carried out in order to detect analogies and differentiations between different eddy current signals. Moreover, cross-correlation between RPs of a reference benchmark and testing eddy current signals has been applied in order to emphasize a different dynamical behaviour and to detect a particular flaw’s shape. In this way, a real-time algorithm for defect shape classification has been performed. Originality/value – Proposed approach is very interesting, and it is an innovation in non-destructive testing procedures. In fact, the shape identification of a flaw is still an open challenge. The proposed approach, based on dynamic analysis, gives the key to solve this particular ill-posed problem, by introducing a relation between the eddy current measurements and the shape of defect existing in the inspected specimen. Very interesting preliminary results have been obtained.