Damage detection based on sparse virtual element boundary measurement using metal-core piezoelectric fiber

Abstract

Pseudo-excitation approach is a recently developed vibration-based damage detection method, exhibiting some appealing features for structural health monitoring applications. However, two main bottlenecking problems, that is, dense measurement points and venerable noise immunity, hamper its use in practical applications. This article tackles these problems by proposing a novel method based on sparse virtual element boundary measurement using metal-core piezoelectric fiber sensors. Different from the local “point-by-point” interrogation modality used in the original pseudo-excitation approach, the proposed method divides the entire structure into several virtual elements to construct a damage location index, describing the damage-induced dynamic perturbation in the corresponding virtual element. To avoid the high-order derivative calculation, which is mainly responsible to the low noise robustness of the original pseudo-excitation approach, metal-core piezoelectric fiber sensors are used to directly measure the surface strains, but only at the virtual element boundaries, leading to a significantly reduced number of measurement points. Experiment is designed and carried out using a cantilever beam, in which a 10-metal-core piezoelectric fiber sensor array is embedded in the structure. Along with the sparse laser Doppler vibrometer measurement, a normalized damage location index is constructed. Results demonstrate that the proposed method not only enhances the noise robustness but also allows a significant reduction in the number of measurement points.