An improved beamforming technique for increased imaging resolution in GW-based SHM

Structural Health Monitoring (SHM) of 2D areas can be performed by combining time waveforms acquired by an array of Guided Wave (GW) transducers through phased addition algorithms. However, spatial resolution in defect localization is often limited by both dispersion in wave mode propagation and imaging noise associated to conventional delay-and-sum (DAS) beamforming. Starting from a statistical interpretation of DAS, this paper combines a recently proposed time-frequency procedure based on the Warped Frequency Transform (WFT) and convex optimization with an adaptive thresholding technique to process multi-mode and dispersive Lamb waves. The proposed signal processing is applied to time waveforms recorded at an array of scan points after waveguide excitation. The results show high-resolution source and scatterer imaging, thus considerably improving localization accuracy by imaging noise removal.