Virtual Frequency Fusion Technique: Detecting Multiple Damages via Broadband Lamb Wave Field Using Scanning Laser Vibrometry

This article presents an efficient frequency fusion method by virtual acquisition of multifrequency guided wave velocity data for robust and accurate localization of multiple damages in plate-like structures. Lamb waves are actuated by applying voltage excitation to piezoelectric patch transducers affixed on the plate, and the velocity response is measured using the noncontact scanning laser Doppler vibrometer (SLDV). The results show that the usual root mean square (RMS) and weighted RMS maps of the velocity field measured using a single narrowband pulse excitation of a given central frequency may not accurately localize multiple damages. To overcome this challenge, we propose a fusion method in the form of the product of normalized reciprocal root mean square ($r\mathrm{RMS}$) maps at multiple narrowband excitation frequencies for combining the measurement data from the different excitation frequencies. However, SLDV measurements at multiple narrowband frequencies would be tedious, time-consuming, and expensive. This difficulty is eliminated by virtually computing the velocity fields at various narrowband frequencies from a single physical measurement using broadband excitation. The efficacy of the proposed method is verified through the experimental investigation on an aluminum plate with multiple damages of different shapes and orientations for two different damage scenarios. The method could identify all damages accurately and clearly through sharp peaks in the $r\mathrm{RMS}$ amplitude at the damage locations without any accompanying ripples near them. On the other hand, the RMS and weighted RMS maps do not perform as well even after applying fusion of multiple frequencies.