LOCATION OF SLIGHT FAULT IN ELECTRIC MACHINE USING TRIGONOMETRIC SPLINE FREQUENCY MODULATION WAVELET TRANSFORMS

Real wavelets can only extract the amplitude of signals. When fault signals change slightly, real wavelets cannot detect the suddenly changing point. Therefore, sometimes it is difficult to detect the occurrence of fault by real wavelets. Using complex wavelets not only the amplitude features of signals can be extracted, but also the phase features of signals, and then the catastrophe point of faint fault signal can be accurately judged by the specific property of phase change under the condition that the amplitude variation of wavelet transform of the signals of faint fault is not evident. Here the frequency modulation wavelet transforms are applied to the diagnosis of faint fault of electrical machines. In this method taking the 2nd order trigonometric spline wavelet as the window function and let q=0, the simplified 2nd order trigonometric spline frequency modulation wavelet transforms (TSFMWT) is obtained. Thereby the phase features of the fault signal is extracted and the catastrophe point of the signals of faint fault can be successfully determined.