In-sensor signal processing techniques for Lamb waves inspections

Abstract

Structural health monitoring based on guided waves (GW) is typically performed by controlling a large number of piezoelectric transducers placed on the component to be inspected. The permanent installation of this technology allows to acquire information about the structural integrity on demand, but it has several limiting factors: bulky hardware instrumentation and large number of connecting cables, complex signal processing, high power consumption and, consequently, high integration costs. Such limitations hamper the adoption of the GW technology in application domains with stringent weight requirements (e.g., aerospace and automotive).These limitations can be addressed by adopting transducers with inherent beam steering capabilities. Such capabilities can be achieved by properly shaping the electrodes of the piezotransducers according to the dispersion characteristics of the waveguide to produce anisotropic wavenumber filtering effects. This “In-Sensor” signal processing can be used to selectively generate ultrasonic waves along arbitrary directions depending on the frequency content of a single excitation signal, or to automatically detect the direction of arrival of mechanical waves propagating within the structure with minimal hardware/software requirements. [In this work (supported by the Horizon2020 project GW4SHM), numerical and experimental validations of this concept will be presented, together with possible future developments.]