Smart classifier based prognostics and health management of induction motor

Abstract

With the utilization of numerous electric engines in everyday life, from transportation and clinical treatment to military action and correspondence, astonishing disappointments can provoke the loss of significant people or an extravagant end in industry. The power, dependability, control, usability, and low upkeep are key highlights to the Induction Motors. In this way, with the expansion in reliance on the machines, it gets crucial to keep up the engines running accurately with no breakdown in light of the fact that a little breakdown can prompt a generous financial misfortune. To accomplish no interruption in measure condition monitoring comes into the image, which chips away at prescient upkeep. In this work, two significantly happening flaws, the broken rotor bar (BRB) and Bearing deficiencies, have been thought of. The current squared or force signal information is obtained from the 1-hp 3-Phase Induction Motor (IM) utilizing the MYDAQ arrangement of National Instruments. Time-Domain (TD) features were extracted from healthy, 1 BRB, 3-BRB, internal race, an external race shortcomings signals. Creating twenty samples of each complete hundred feature vector, which is utilized for training, testing, and validation. After normalization, this data set is used for the comparison of Ensemble and Decision Tree smart classifiers based on model accuracy. It is found that because of the heterogeneous type classification approach the Ensemble achieved maximum accuracy than the decision tree of 95.8%.