Numerical and experimental investigation of delamination severity estimation using local vibration techniques.

This paper presents vibration-based methods for detecting and quantifying delaminations in composite structures. Three different delamination sizes were considered numerically and experimentally. Our approach combines and compares finite element method (FEM) numerical analysis and experimental measurements to find the location of structural delaminations and identify their severity. First, we noticed that a low-frequency modal study comparing pristine and damaged plates could not provide local damage information. Hence, a high frequency-modal analysis was performed to obtain the local damage information. The damaged mode shapes corresponding to resonance frequencies that cause significant local vibration (high amplitude) over the damaged regions were calculated numerically. The result shows the capability of damage identification (location and size of structural delaminations) within the high-frequency modal analysis. The numerical results were validated experimentally. The experiments include electromechanical impedance spectroscopy (EMIS) and scanning laser Doppler vibrometer (SLDV) techniques. The EMIS experiment was implemented to predict the resonance frequencies which cause significant local vibration over the delamination regions. The spectra of the electromechanical response resemble the vibration response of the structure and detect the new resonances created by the presence of the delamination. The test plate's experimental resonance peaks show a good match with the resonance frequencies of the numerical modal analysis. The SLDV experiments measured the local vibration over the investigated scanning areas. The operational vibration shapes corresponding to predicted resonance frequencies were created. These experimental mode shapes may visually measure the delaminations by observing intense local vibration over the delamination zones. Comparing the numerical mode shapes and the operational vibration shapes shows a good match between numerical and experimental results. [ABSTRACT FROM AUTHOR]