Reference-Free Fatigue Crack Detection Using Nonlinear Wave Modulation

Abstract

This study presents a reference-free fatigue crack detection technique using nonlinear wave modulation. Two independent input signals are generated using two surface-mounted lead zirconate titanate (PZT) transducers; a low frequency (LF) signal and a high frequency (HF) signal. The presence of a structural nonlinearity, such as a fatigue crack, can provide a mechanism for nonlinear wave modulation, and create spectral sidebands around the frequency of the HF signal. A signal processing technique combining linear response subtraction (LRS), synchronous demodulation (SD) and continuous wavelet transform (CWT) filtering is developed to extract only the crack-induced spectral sidebands. Actual fatigue crack is detected without any data obtained from intact condition is developed using sequential outlier analysis. Finally, the robustness of the proposed technique is demonstrated under temperature and loading variations at two distinct sensing locations.