Localisation and quantification of micro-scale corrosion on aluminium plates based on spatial dissimilarity analysis of Lamb waves

Corrosion monitoring is essential for assessing the safety and longevity of metallic structures. To address challenges related to feature extraction in signal processing and the limited detection capability for incipient corrosion damage on metal plates, this work presents a novel method for localising and quantifying micro-scale corrosion based on the spatial dissimilarity analysis of Lamb waves. Initially, wavelet transform is utilised on the received signal, isolating the S0 wave packet through variations in wavelet coefficients. The Gaussian Radial Basis Function (RBF) matrix is then computed to reveal the spatial dissimilarity of the segmented signal, effectively capturing its subtle nonlinear characteristics. Singular Value Decomposition (SVD) is subsequently applied to extract key features from the obtained similarity matrix. Finally, the Discrepancy Index (DI) between baseline and test signals is defined by the maximum singular value of the Gaussian RBF matrix derived from the damage signal, facilitating precise localisation and quantification of corrosion damage on the plate. Corrosion experiments are conducted on an aluminium plate to verify the effectiveness of this method. The results demonstrate its robustness and accuracy in locating and quantifying micro-scale corrosion damage, underscoring the reliability and high precision of this approach for micro-scale corrosion assessment in metallic structures.