Contact-type defects are prevalent in composite constructions and sandwich panels due to failure mechanisms such as bolt loosening and delamination. Contact acoustic nonlinearity is a manifestation of such defects that nonlinear health monitoring systems can detect. Vibro-acoustic modulation (VAM) is a well-established technique for the early detection of nonlinear defects in structures. It employs bi-tone excitation to reveal damage-induced nonlinearity, which results in the appearance of sidebands in the response spectrum. The objective of this research is to use PZT-based excitation to monitor bolt loosening in a sandwich beam in real-time. To address the limited capacity of the PZT transducers to excite the nonlinear mechanism, a sensitivity analysis (SA) for input factors of VAM testing was conducted to improve damage detection using the results. The Morris approach is used to investigate the sensitivity of VAM damage metrics derived analytically in the frequency domain. To reduce the number of tests required for the experimental SA, the response surface methodology (RSM) is applied. The analysis of variance and Fischer's statistics are used to calculate the SA of experimental damage indices. In RSM computations, the discrete influence of excitation frequencies on the modulation sidebands is addressed. For reliable SA of experimental results, the effect of background noise is taken into account. The findings of this study may be applied to the selection of appropriate damage metrics in real-world applications of VAM health monitoring systems, as well as the effective tuning of input control parameters to maximize damage detectability.
Read full abstract