A benchmark for vibration-based structural damage assessment of composites using wavelet analysis

The presented benchmark is a result of 5-year simulation and experimental studies on damage assessment in composite structures. In a group of cases the damage was modelled using finite element analysis software, while the other group of results was obtained during bench tests. The benchmark consists of 143 problems of damage detection, localization and identification in beam-type and plate-type structures with various types of damage including cracks and notches, surface damage, delamination, impact damage etc. Applied algorithms use the modal shapes of vibration of tested structures with damage as an input, and using developed wavelet-based methods the diagnostic information about the damage presence, its location and type is obtained during the analysis. Numerous methods based on various types of wavelet transforms are available for the analysis of damage presence and location basing on modal shapes of vibration. They include classical wavelet transforms (continuous, discrete, stationary and undecimated ones) as well as more advanced transforms like fractional discrete wavelet transform or quincunx wavelet transform. Some of these algorithms are combined with optimization procedures, which allow to select optimal parameters of wavelets in order to improve detectability of damage. Several additional benchmark problems consist of 3D problems obtained from computed tomography. The wavelet-based algorithms are applied there in order to detect, localize and classify internal defects in composite structures. The graphical user interface in Matlab environment was provided for testing purposes of the benchmark.