Font Size: 

Time-frequency representations of vertical transient responses of a rotor are investigated for fatigue crack detection. The rotor is simply modeled as a Jeffcott rotor to obtain transient responses in vertical direction. In a rotating shaft, the crack opens and closes during each revolution of the shaft, which is considered as a breathing behavior. In this study, the breathing behavior of the crack is modeled using a proposed explicit softly-clipped cosine function. The breathing crack excites super-harmonic components of the vertical transient responses, and consequently make changes in their time-frequency representations obtained from Hilbert-Huang transform based on an improved empirical mode decomposition (EMD) algorithm. The improved EMD algorithm employs piecewise cubic Hermite interpolating polynomial (PCHIP) instead of cubic spline interpolation used in EMD. Using PCHIP eliminates the extremum interpolation problem in empirical mode decomposition algorithm, and provides more feature information in time-frequency representations of transient responses.