Motor current signature analysis through machine learning techniques
Diya Dharmendra, S. Suchitra · 2024
This article discusses implementing a machine-learning model to classify motor faults using its current signature. Rotating electrical machines, like induction motors, are important in the electrical industry. It has a simple construction, increased reliability, and easier maintenance. Induction motors are prone to numerous faults like stator and rotor faults, rotor eccentricity, winding, and bearing faults during their operation. The thing that plays a significant role in detecting these incipient motor faults is the motor current. Therefore, for identifying these typical machine problems, Motor Current Signature Analysis (MCSA) is thought to be the most widely used and reliable fault detection technique. Because the induction machine is nearly symmetrical, changes in the flux interaction between the rotor and stator lead to changes to the stator&s;s voltages, current, vibration, and emf. The current signature recorded with the base as time is converted to frequency base using the FFT algorithm and different classification models were used to classify each fault according to its current signature.