Damage identification in aluminum plates based on iterative partition algorithm using waveform centroid

In the field of structural damage monitoring, sensor arrays are often used to detect damage sites. However, the use of a large number of sensors increases the required amount of data processing, which imposes severe restrictions on effective online monitoring of industrial applications. In this study, an area-division method based on a waveform centroid algorithm is proposed to locate damage sites accurately using a limited number of sensors. Experiments are conducted on a 600 mm x 600 mm x 1.5 mm aluminum alloy plate to evaluate the performance of the proposed method. Four piezoceramics, acting as exciters and receivers, are arranged on the central region of the plate, dividing the entire detection area into eight portions. In order to scan the plate structure, Lamb waves are generated from each of the piezoceramics while the others function as receivers. By performing three successive comparisons after processing the data corresponding to each exciter-receiver pair via the waveform centroid algorithm, the specific portion containing the damage site can be identified. Subsequently, two ellipses are drawn using the three piezoceramics neighboring the identified portion, to visually locate the damaged site. The technique is applicable under conditions with low signal-to-noise ratios with an error of 1.5% in this study. It has the potential to be useful in industrial applications, in the context of real-time structural health monitoring. (C) 2021 Elsevier B.V. All rights reserved.