Structural damage detection in composite aeronautical components through methods based on digital filtering and wavelets transforms

Several recent investigations have shown how visual inspections of mode shapes do not ensure an accurate, or even rough, identification of structural defects. This paper aims at exploring the applicability of a numerical-experimental method based on digital shape processing techniques for the identification of localised structural changes. The problem is tackled within an aeronautical context where the damage detection procedure is specifically applied to multilayered plates having symmetrically stacked layers. The examined components are used in the aeronautical industry during the manufacturing process of aircrafts and being, such structural components, an essential part for the integrity of the structure, they undergo extremely accurate investigations before becoming a part of the whole system. An interesting point here investigated is the possibility of offering a less demanding NDT alternative to those existing (i.e. ultrasonic). Several defects are simultaneously introduced into multilayered plates; these defects are simulated through a finite element model basing the model on a template derived from an ultrasonic benchmark. In order to evaluate the influence of real measurements in the identification of localised defects, noise has been artificially introduced in the model to simulate a reduced signal-to-noise ratio in accordance with currently available laser measuring technology. The capability of the presented techniques of identifying localised defects is compared with the ultrasonic benchmark and relevant conclusions are extracted.