Identification and characterization of damaged fiber-reinforced laminates in a Bayesian framework

Non-destructive testing of damaged composite laminates modeled from the homogenization of fiber-reinforced polymers has always been and still is a challenge, both by underlying complexity and the difficulties encountered in the quantification of uncertainties related to the identification and characterization of defects. In order to provide a rigorous framework that accepts data from different modalities and allows data fusion as well, we propose here a Bayesian neural network (BNN) with two input streams, with a focus on local inter-layer delaminations identification and characterization.