Quantitative damage assessment using guided ultrasonic waves signals

摘要

Inverse algorithms based on neural network were developed for quantitative assessment of damage in structures, e.g. cracks in aluminum plates and delamination in laminate composite beams or plates. It uses a concept of digital damage fingerprints (DDFs) extracted from scattered wave signals, which are the input for training of artificial neural network. The trained neural network was used for inverse assessment of the damage, and the algorithm was validated by experiments with actual damage introduced in aluminum plates and laminate composite beams or plates, where DDFs were extracted from networks of piezoelectric elements (PZT) attached to the surface of the structures for activating and capturing of guided ultrasonic wave signals. The results predicted by the algorithm show good accuracy in defining damage parameters, such as central position, size, orientation etc., for cracks and delamination in the structures.