FAULT DIAGNOSIS OF A BRAOKEN ROTOR BAR IN AN INDUCTION MOTOR USING MOTOR CURRENT SIGNATURE ANALYSIS

Abstract

The project report entitled "Fault Diagnosis of a Broken Rotor Bar in an Induction Motor using Motor Current Signature Analysis” reports about Motor Current Signature Analysis which is a non-invasive and cost-effective method for the fault diagnosis and classification of induction motors. This technique uses the current signals generated by the induction motor during its operation to diagnose faults such as rotor and stator winding faults, bearing faults, and misalignment faults. The process involves capturing the current waveform of the induction motor, processing it using signal processing algorithms, and analyzing it to identify the fault. The method is based on the principle that different faults generate distinctive current signatures. The current signals generated by the healthy and faulty induction motor, loaded with different loads, is captured using a current transformer, which is connected to a data acquisition system MyDAQ. The captured data is then processed using signal processing algorithms which is then processed to extract relevant features that are used to train machine learning models. The result indicates that Naïve Bayes algorithm was able to classify health condition of an induction motor with the accuracy of 94.4%. The algorithms like SVM and Decision tree also performed well with an accuracy of 88.9% and 91.7% respectively. The trained models can then be used to perform real-time fault classification on the induction motor, providing valuable information for predictive maintenance and condition monitoring