Temporary short circuit detection in induction motor winding using second level haar-wavelet transform

Faults in induction motor winding can be successfully detected using different motor current signature analysis. However, there still remain some parts where the performance of conventional methods can be improved. In case of the fast Fourier transform (FFT) method, it can only identify the permanent fault, but not the temporary one because the method gives frequency content similar to the normal condition. Moreover, the FFT technique is unable to provide the exact timing information of the fault occurrence. On the other hand, the method based on the first level wavelet transform sometimes gives misleading information, especially in case of starting and ending points of temporary short circuit. For these reasons, this paper comes up with a new method for winding fault detection, which analyzes motor current spectrogram based on extension wavelet analysis, called the second level Haar wavelet transform. The proposed method is able to detect temporary fault with very short duration and low current level with more clear information than that of the first level. Several testing scenarios are presented to confirm the robustness of the proposed method including the provision of time of occurrence information for each case.