Application of TFD tools for the tracing of eccentricity-related components in induction machines

The analysis of the stator startup current for the diagnosis of mixed eccentricities has been revealed as an alternative reliable way for detecting the presence of this fault. It avoids some of the constraints of the classical diagnosis approach, based on the identification of the eccentricity-related components appearing in the FFT spectrum of the steady-state current. In this context, the selection of suitable Time-Frequency Decomposition (TFD) tools for the analysis of non-stationary quantities-such as the startup current- plays a crucial role. Previous works have proven the reliability of the discrete wavelet transform (DWT) and Hilbert-Huang transform (HHT) for tracing the transient evolution of eccentricity-related components and, therefore, for the diagnosis of the fault. In this paper, a different tool, the Wigner Ville distribution (WVD) is applied to the diagnosis of mixed eccentricities, making special emphasis in the detection of high frequency components introduced by this fault. Moreover, a brief comparison with respect to HHT and DWT is carried out, showing the main advantages and drawbacks of each tool regarding the transient-based diagnosis of the aforementioned fault.