Development of nonlinear spectral correlation between ultrasonic modulation components

Abstract

When a damaged structure is exposed to ultrasonic waves at two distinct frequencies, nonlinear wave modulation resulted from a damage such as a fatigue crack can be observed in the corresponding response, offering an opportunity for early damage detection. This study attempted to improve nonlinear wave modulation-based damage detection by applying statistically weak-linked inputs over two distinct frequencies to the target structure and then conducting a cyclostationarity analysis of the corresponding structural response. The cyclostationary nature of the structural response produces a statistical variation over time, allowing the structural response to be processed using a spectral correlation function. The spectral correlation can thus be used to transform the structural response into two dimensions characterized by their cyclic and spectral frequencies. Then, the damage-induced nonlinear modulation can be detected by studying the spectral correlation values for the specific cyclic and spectral frequencies defined by the modulation frequencies. This premise forms the basis for nonlinear spectral correlation, which is a new damage feature that is superior in terms of its sensitivity for nonlinear damage and improved robustness against noise compared to a conventional spectral density function. The performance of the proposed technique was validated by conducting an experiment with aluminum plates containing real fatigue cracks. The results of this experiment showed that the proposed technique could detect damage even under severely noisy conditions.