Numerical and experimental assessment of the modal curvature method for damage detection in plate structures

Use of modal curvatures obtained from modal displacement data for damage detection in isotropic and composite laminated plates is addressed through numerical examples and experimental tests. Numerical simulations are carried out employing COMSOL Multiphysics as finite element solver of the equations governing the Mindlin-Reissner plate model. Damages are introduced as localized non-smooth variations of the bending stiffness of the baseline (healthy) configuration. Experiments are also performed on steel and aluminum plates using scanning laser vibrometry. The obtained results confirm that use of the central difference method to compute modal curvatures greatly amplifies the measurement errors and its application leads to unreliable predictions for damage detection, even after denoising. Therefore, specialized ad hoc numerical techniques must be suitably implemented to enable structural health monitoring via modal curvature changes. In this study, the Savitzky-Golay filter (also referred to as least-square smoothing filter) is considered for the numerical differentiation of noisy data. Numerical and experimental results show that this filter is effective for the reliable computation of modal curvature changes in plate structures due to defects and/or damages.