Development of a quantitative measurement procedure for defect size and morphological characterization in composite materials by Shearography

In the last decades, new materials and solutions were studied in the field of aerospace industry in order to reach weight reduction of structures, reliability, durability and mechanical strength. In this scenario, composite materials play an important role and the study of suitable Non-Destructive-Testing (NDT) procedures is crucial. In this dissertation, the focus is put on composite materials inspection by using an optical, interferometric technique, Shearography, whose main application is defect and damage detection in composite materials. The use of Shearography is still qualitative: the user is only capable of stating if a defect is present or if it is not, by observing typical interference fringes appearing in case of defect. Shearography, in fact, measures the first derivative of the out-of-plane displacement of an observed surface, by considering variations in the phase relation between separated points of the surface itself. A fringe pattern, i.e. a variation in the phase relation between these points is detected if, and only if, a differential deformation occurs between different regions of the surface. This is what happens when a structure is damaged: the defect region of influence will be more warped with respect to the rest of the structure. The need of observing deformations implies the inspected structure to be loaded. Three kinds of load can be exploited for Shearography inspection: mechanical, vacuum and thermal load. Thermal load implies a series of effects difficult to forecast and/or simulate, such as environmental absorption, air turbulences, propagation of the radiation inside the material in cases where material properties are unknown, etc. For all these reasons, no standard procedures were given even now. In addition, Shearography performances and results are highly dependent on measurement conditions and parameters. The first influencer parameter is the shear distance, i.e. the distance between points used for evaluating phase relations, the second one is the distance between the inspected surface and the Shearocamera, affecting spatial resolution, the third one is the load to apply to the structure: the higher the load, the higher the overestimation. On the other hand, the lower the load, the lower the Signal-to-Noise Ratio (SNR). Other influencer conditions are the defect characteristics themselves, i.e. the defect depth and size, whose combination affects the needed load and the sensitivity required. Therefore, the aim of this work is to provide a deep study of Shearography technique in order to overcome its two great limits, i.e. the lack of algorithms capable of accurately determining defect size and morphology, and the lack of a repeatable procedure for thermal load. The first goal was reached by designing an algorithm that studies the phase maps involving both image analysis tools and signal processing ones. This was fundamental since it was demonstrated that analyzing the phase maps only by standard image processing methods for morphological study does not lead to good quantitative estimations neither for simple nor for complex damages. Consequently, a series of phase profiles to be trated as pure signals are extracted each time a phase map has to be processed. At this point, the defect shape is reconstructed by localizing the boundaries of each phase profile. This was obtained by the use of Wavelet Transform method for singularity estimation, in addition with another function, called Structural Intensity, that enables to select only the dominant singularities. Finally, the second goal was achieved by two different steps. At first, a synchronized and automatic test bench was designed and prototyped, making the acquisition repeatable and reducing uncertainty sources. The second step lies in the elaboration of all the acquired data corresponding to different combinations of parameters, that are matched together in order to define the optimal condition.