Wavelet feature analysis based induction motor internal fault identification using experimental data

Internal faults in induction motors such as bearing faults, eccentricity faults and stator faults are necessary to detect at early stages because if left undetected they can develop further into serious mechanical faults. To detect these faults at early st age, when they are developing a wavelet transform based approach is proposed in this work. The stator current of motor is decomposed to various level using discrete wavelet transform to isolate specific frequency bands associated with fault characteristics. From these decomposed signals, wavelet energy, entropy and peak amplitude features are extracted. The presence of fault conditions, it’s type and severity are described after accessing the extracted features. Experimental data for healthy, outer race bearing fault, static and dynamic eccentricity fault and stator interturn fault are analyzed using the proposed approach. Analysis of results demonstrate the effectiveness of proposed method, emphasizing it’s potential to use for improving the dependability of induction motor by providing a non-invasive and effective diagnostic technique.