Smart classifiers based classification and condition monitoring of induction motor faults

Abstract

With the advancements in the field of automation in the industries the use of machines is very high and if the machine which require some rotatory action for the load, the Induction motor comes in to play because of the advantages such as robustness, low maintenance, low cost etc. But with the increase in the dependency over the motors it becomes highly recommended to have machines with reliability because break in the work can lead to huge amount of loss. In order to increase the reliability of the motors predictive maintenance comes into play which requires fault classification or detection which is easily and accurately possible using the Machine Learning algorithms. With the requirements of the present scenario for predictive maintenance, this paper presents the fault classification of induction motor using Support Vector Machine SVM) and K-Nearest Neighbour (KNN) technique of classification. Here in this paper the bearing fault (BF) and broken rotor bar (BRB) fault is considered. The results collected are on the basis of validation and Principle Component Analysis (PCA) technique. And it is found that the SVM technique is better than the KNN for fault classification of Induction motor.