Lamb wave frequency–wavenumber analysis and decomposition

Abstract

Lamb waves have shown great potentials in damage detection of thin-walled structures due to their long propagation capability and sensitivity to a variety of damage types. However, their practical adoption has been hindered due to the complexity caused by their multimodal nature. Various wave modes that propagate at various velocities in the structure make the interpretation of Lamb wave signals very difficult. It is desired that the modes can be separated for independent analysis or further employment. In this article, we present our studies on the multimodal Lamb wave propagation and wave mode decomposition using frequency–wavenumber analysis. Wave representation in the frequency–wavenumber domain is obtained using multidimensional Fourier transform, where various Lamb wave modes can be easily discerned. This allows for separating them or extracting a desired wave mode through a filtering process, thus making it possible to use a single-mode Lamb wave for the detection of a certain type of damage in structural health monitoring applications. To retain the temporal and spatial information that is lost during Fourier transformation, a novel wavenumber analysis is also presented. These concepts are illustrated through experimental testing where high spatial resolution wavefields are measured by a scanning laser Doppler vibrometer.