Session: 04-15-01: Congress-Wide Symposium on NDE & SHM – NDE and Prognostics in Structural Applications

Paper Number: 68635

Start Time: Tuesday, 10:05 AM

68635 - Spectral Correlation Method for Fatigue Crack Detection Based on Nonlinearly Mixed Ultrasonic Waves 

The recent development in a nonlinear ultrasonic two-wave mixing technique has made it possible to detect fatigue crack at its early-stage. This technique introduces two input waves into a nonlinear structure (e.g., fatigue crack) at distinct frequencies. The interaction of these waves with each other, gives rise to nonlinear second-order combined harmonic (SOCH) components, causing them at sum and difference of input components. However, the amplitudes of the SOCH components are generally one or two orders of magnitude smaller than that of the input linear components. Consequently, the amplitude of the SOCH components generated by the fatigue crack gets considerably feeble in a noisy environment. To effectively overcome this challenge, the conventional linear power spectrum and bispectrum methods are widely been used to distinguish nonlinear SOCH components, which result from fatigue crack and to remove those, which do not. The major restriction in the conventional methods is that they are mainly meant for stationary signals and are not adequately defined for nonstationary signals. This study presents the collinear two-wave mixing technique with a spectral correlation method as a better alternative to enhance the performance of nonlinear SOCH components. The proposed spectral correlation considers not only nonlinear SOCH components in ultrasonic waves but also a spectral correlation between the nonlinear SOCH components.

The uniqueness of the proposed technique lies in the followings: (1) the present study uses only a single transducer for simultaneous generation of two input waves instead of using two individual transducers, (2) the spectral correlation of nonlinear SOCH components is insensitive to noise interferences, (3) the tone burst inputs are used to the contrast of spectral correlation between intact and fatigue crack conditions rather than the sinusoidal inputs, (4) extracted nonlinear SOCH components from the spectral correlation method are more reliable than the obtained nonlinear SOCH components using the conventional methods,  and (4) the proposed approach can be successfully applied to fatigue crack detection.

A theoretical model is presented to obtain the relationship between the spectral correlation coefficient and nonlinear SOCH components. Experiments are conducted on aluminum plate specimens with fatigue crack to validate the performance of the spectral correlation. The experimental results demonstrate that the extracted nonlinear SOCH components prove to be sensitive to fatigue crack. The noise contents in the original signal are completely removed by extracting the nonlinear SOCH components using the spectral correlation method. Furthermore, the estimated fatigue crack length from the proposed approach correlates well with the actual fatigue crack length.

Presenting Author: Santhakumar Sampath KAIST

Authors:

Santhakumar Sampath KAIST
Hoon Sohn KAIST