Abstract
•Investigation of the wave propagation features in a honeycomb sandwich plate over a broadband frequency range;•Robust experiment-based identification process for estimating propagation direction-dependent and frequency-dependent wavenumber-space characteristics and mechanical properties in presence of uncertainties;•Statistical investigations of the impact of measurement variability on the identified parameters;•Generalized Polynomial Chaos method for low-cost robust identification.
The purpose of the present paper is to investigate the wave propagation features in a sandwich plate with honeycomb core over a large frequency domain. Robust experiment-based identification processes are proposed to estimate propagation direction-dependent and frequency-dependent wavenumber-space (k-space) characteristics in presence of parametric uncertainties. These processes combine structural identification and uncertainty propagation methods. A special emphasis is put on wave correlation methods compared to two-dimensional Discrete Fourier Transform. The Variant of the Inhomogeneous Wave Correlation method is extended here to two-dimensional identification problems. The vibration field is experimentally measured at points which geometric coordinates are supposed to vary randomly. Statistical investigations are then carried out to quantify the impact of the measurement points geometric coordinates’ variability on the identified parameters and evaluate the robustness of the proposed identification processes against uncertainties. Valuable insights into k-space profiles, damping loss factor and equivalent mechanical parameters, regarding the structural orthotropic behavior, are highlighted. The obtained results show the large variability of the identified parameters and reveal a significant identification sensitivity to the measurement points geometric coordinates’ uncertainties. The use of the generalized Polynomial Chaos method allows robust identification with an interesting computing time reduction regarding the Latin Hypercube Sampling.