Spectral Correlation-Based Enhanced Breathing Crack Diagnosis of Steel Structures with Linear Response Subtraction Scheme

In this paper, a reference-free spectral correlation-based approach for enhanced breathing crack diagnosis is proposed using the acceleration time history responses measured under bitone harmonic excitation. Even though spectral correlation can effectively remove the stationary noise components overlapped with the nonlinear intermodulation components in the response, it is difficult to isolate the intermodulation components alone at a selected cyclic frequency, and careful choice of appropriate cyclic frequency has to be made for accurate damage diagnosis. Therefore, linear response subtraction is first applied on the actual response to isolate the breathing crack induced nonlinear intermodulations, and then, spectral correlation computed from modified response is used for breathing crack diagnosis. The proposed linear response scheme helps in enhanced breathing crack diagnosis using spectral correlation estimated at any cyclic frequency. The proposed scheme is verified through both numerical and experimental examples.