Application of the Novel Wavelet Ann Method for Segregating Bearing Faults in Three-phase Induction Motor

Induction motors are a critical component of many industrial processes and are frequently integrated into commercially available equipment. Safety, reliability, efficiency, and performance are some of the major concerns of induction motors.

Three-phase induction motors are the ‘workhorses’ of industry and are the most widely used electrical machines. For this reason, the detection of motor failures is very important. Bearing problems are one of the major causes of drive failures. Early detection of bearing faults allows replacements of the bearings rather than replacement of the motor. The present contribution reports experimental results for monitoring bearing faults in induction motors. Motor line currents have been analyzed using modern signal processing and data reduction tools combining Park’s Transformation and Discrete Wavelet Transform (DWT). Feed Forward Artificial Neural (FFANN) based data classification tool is used for fault characterization based on DWT features extracted from Park’s Current Vector Pattern. An online algorithm is tested successfully on a three-phase induction motor and experimental results are presented to demonstrate the effectiveness of the proposed method which can reliably distinguish the inner race and outer race defects of the bearing. It is observed that for five processing elements in the hidden layer, 100% classification accuracy is achieved for healthy and faulty conditions.

Experimental results are presented to demonstrate the effectiveness of the proposed methodology. The study concluded that the proposed methodology is useful in identifying inner race and outer race defects of bearings in induction motors with 100 percent accuracy.