DFA and DWT based severity detection and discrimination of induction motor stator winding short circuit fault from incipient insulation failure

Modern research surveys emphasize that stator winding fault holds a significant percentage in induction motor failures. In Motor Current Signature Analysis (MCSA) based stator winding fault diagnosis, it has been found to be a challenging task to discriminate inter turn short circuit faults from inter-turn incipient insulation failures providing unbalances in three phase motor currents identical to short circuit faults. This paper proposes an approach to achieve this objective by applying Discrete Wavelet Transform (DWT) as signal decomposition tool on the faulty phase motor current captured through CRIO-9075 integrated controller and chassis system having 400MHz power PC controller, LX 25 Gate FPGA with NI 9227. Inverse Discrete Wavelet Transform (IDWT) as signal reconstruction tool has been employed to extract relevant frequency band which is sensitive to stator winding faults. Reconstructed or filtered currents under different fault cases containing specific band of frequency signals are then fed to Detrended Fluctuation Analysis (DFA) algorithm which has the competency to assess trend of fluctuations present in signals under consideration. Results of DFA in the form of short term fluctuation coefficient (αs) and long term fluctuation coefficient (αl) are found to be capable in discriminating the inter-turn short circuit fault from incipient insulation failure. Proposed method is also capable in detecting the severity levels of two different types of fault cases. Entire analysis reported in this work is based on experimentally obtained motor current signals.