Evaluation of Structural Integrity of Metal Plates by Fuzzy Similarities of Eddy Currents Representation

In this paper, we present a practical application of the methodologies introduced by Jaime Gil-Aluja and Lotfi Zadeh in Civil and Electrical Engineering. Metallic plates bi-axially loaded deform producing dangerous mechanical stresses that are not visually appreciable. Being the representation construction of such stress conditions by 2D images extremely complex, in this work, we propose to generate suitable Eddy Currents (ECs) images to translate the information content of mechanical stresses into representative electric signals easier to image. By grouping the produced images in different classes related to different bi-axial loads and in a single class all the images referring to plates in absence of loads, the evaluation of the integrity of a plate is transformed into a problem of classification/decision. This further step is carried out by means of the measure of Fuzzy Similarities (Ss) between the 2DEC signal at hand and the prototypical classes. The achieved performance are comparable to more established approaches that are commonly plagued by a higher computational load. The proposed methodology is also shown to be able to manage uncertainty in an application of relevant industrial interest.