Detection of matrix cracking in composites using Lamb wave propagation and machine learning: numerical modelling and experimental validation

The purpose of this study was to propose an intelligent nondestructive method for the detection of matrix cracking in composites using the Lamb wave propagation and machine learning approaches. First, pure resin and cross-ply laminated composite specimens with different crack densities were fabricated. The Lamb wave propagation experiments were carried out on the fabricated specimens and the phase velocity and amplitude decay of Lamb wave were estimated. Then, Finite element (FE) models were developed to simulate the Lamb wave propagation in the laminated composite specimens. The results obtained from the FE models were compared and found to be in good agreement with those from the experiments. It was revealed that the phase velocity and amplitude decay of the Lamb wave are sensitive to the severity of matrix cracking. The Lamb wave features extracted from the FE models were used to train a supervised multilayer perceptron neural network model for the prediction of crack density. The results indicated that the proposed combined Lamb wave propagation and machine learning approach can be effectively used for the assessment of the matrix cracking severity in polymeric laminated composites.

کلیدواژه ها:

Polymeric composites ، Matrix cracking ، Guided Lamb wave ، Finite element modeling ، Machine learning