Generation of Mode-Selective Frequency Response Functions from Temporally Sampled, Broadband and Multimodal Guided Ultrasonic Wave Signals

Abstract

Structural health monitoring (SHM) systems for thin-walled, large-scale structures using guided ultrasonic waves (GUW) have recently been complemented by methods based on frequency response functions (FRF). These approaches enable the calculation of the structure’s response to any artificially generated excitation signal within the frequency range under investigation. This study presents a novel concept for extracting separate FRFs for the two fundamental modes from temporally sampled, broadband and multimodal GUW signals. In this work, quasi-ideal impulse response functions (IRF) are generated by retransforming multimodal broadband FRFs into the time domain. These FRFs are recorded energy-efficiently using sweep excitation and contain all occurring GUW components at the sensing position. The extended dispersion based frequency-domain intrinsic component decomposition (DBFICD) algorithm and the iterative frequency-domain envelopetracking filter (IFETF) algorithm are used to extract the corresponding signal components from the IRFs. Based on their different dispersive behaviour they can be assigned to the fundamental modes and combined to form the two respective IRFs. Transforming the latter back into the frequency domain results in two separate FRFs. This allows damage detection algorithms to estimate the structure’s mode-selective response to any selected signal within the frequency range under investigation. The method is validated experimentally by recording multimodal and single-mode FRFs in an aluminium plate with a mode-selective, two-sided piezoelectric actuator setup and a laser Doppler vibrometer. The estimated single-mode structural responses based on the FRFs extracted by the presented approach agree well with the actual responses to mode-selective excitation. The presented concept enables mode-selective FRF-based SHM systems to work with multimodal excitation, which reduces setup complexity and increases the application range by providing information for both fundamental modes over a wide frequency range.