Impact Damage Characterization in Composites Using Signal Processing Techniques

Anticipate and characterize damages in layered carbon fiber-reinforced polymers (CFRP) is a challenging problem. Non-destructive evaluation using ultrasonic signals is a well-established method to obtain physically relevant parameters to characterize damages in isotropic homogeneous materials. However, due to their structural complexity, composites require special care in signal interpretation. Some enhancements on the inversion and interpretation are proposed by adapting classical signal processing techniques, including spectral analysis, homomorphic transformations, linear prediction, and by defining weighted residue in the inverse procedure scheme. This study aims to evaluate the feasibility of the signal-processing based inverse problem to reconstruct experimental data sets at different levels of damage, and to present a comparative study between the above-mentioned techniques.