Composite Elastic Property Identification Through Guided Wave Inversion And Damage Detection By Data-Driven Process

The elastic property identification of composite materials is of great interest for the conditional monitoring of lightweight structures, such as modern aircrafts (e.g., B737, A380, etc.). Because of the uncertainties in manufacturing processes for composite materials, it is important to develop a way to measure the properties of the final part in a manner that is non-destructive and can be applied in-situ on an existing structure.The first part of this paper examines the possibility of property characterization of quasi-isotropic multi-layered composite laminates using pseudo-experimental data (phase velocity dispersion curves with true elastic constants). An inversion scheme is proposed by minimizing the discrepancy between phase velocity dispersion curves for the 'true' elastic constants and those for "trial" constants, determined from the Semi-Analytical Finite Element (SAFE) method. Of interest are the three fundamental Lamb modes (the axial mode -S-0, the flexural mode-A(0), and the shear horizontal mode -SH0). A simulated annealing optimization (SA) is exploited to fast locate the global optimum with the smallest discrepancy. The second part of this paper implements data-driven matched field processing to locate the defects on the same type of composite with complicated geometry, including skin and stringer. Using a full matrix data acquisition system, two defect locations are investigated, one is on the skin, the other is the stringer.