Time-frequency distribution decomposition with application to detection of rotor rub-impact fault

In this paper, we introduce a new data analysis method, called time-frequency distribution decomposition (TFDD), to extract the components of vibration signal one by one for rotor rub-impact fault. When the early rub-impact fault occurs in the rotor system, the vibration signal will present frequency modulation feature because of the periodic rub-impact between the stator and the rotor. Through approximating phase and amplitude functions of each component of vibration signal, the feature of the original signal can be characterized by these functions. Moreover, each component can be analyzed individually. The validity of the method is demonstrated on a real rotor system of a gas turbine with rub-impact fault. The analysis of the application shows that the TFDD is powerful in analysis of nonlinear and nonstationary signals and is an effective tool for the detection of rub-impact faults.