Lamb wave-based quantitative identification of delamination in CF/EP composite structures using artificial neural algorithm

Abstract

Delamination in composite structures plays a major role in lowering structural strength and stiffness, consequently downgrading system integrity and reliability. A Lamb wave-based quantitative identification technique for delamination in CF/EP composite structures was established. Propagation of Lamb waves in a series of composite laminates, individually bearing a delamination, was evaluated using dynamic FEM analyses. Taking advantage of wavelet transform and artificial neural algorithms, an Intelligent Signal Processing and Pattern Recognition (ISPPR) package was developed, by which the spectrographic characteristics of simulated Lamb wave signals in the time-frequency domain were extracted and digitised as Digital Damage Fingerprints (DDF), to construct a Damage Parameters Database (DPD). The DPD was then used offline to train a multi-layer feedforward artificial neural network (ANN) under supervision of an error-backpropagation (BP) algorithm. Assisted by an active online structural health monitoring (AO-SHM) system with an active piezoelectric actuator/sensor network, the proposed methodology was validated online by identifying actual delaminations in CF/EP (T650/F584) quasi-isotropic composite laminates.