Rotor broken bar diagnostics in induction motor drive using Wavelet packet transform and ANFIS classification

The paper proposes diagnostic technique for identifying rotor broken bar in speed sensorless three-phase induction motor drive system. Experimental data is collected for healthy and faulty motor with two rotor broken bar. The collected data is first processed using Wavelet packet 1-D transformation to extract the statistical information. The statistical information thus obtained is used to train the Adaptive Neuro-Fuzzy inference system (ANFIS) which is used as fault classifier. The stator current space vector magnitude is used in conjunction with Wavelet transform as the current signature is the most effective method of fault diagnosis. The speed signal obtained from the observer system is directly used in ANFIS. The proposed synergy of Wavelet packet and ANFIS provide highly accurate and computationally efficient tool that can be used for on-line fault diagnosis.