Broken Rotor Bar Fault Detection Using Advanced IM Model and Artificial Intelligence Approach

Abstract

In this paper, a reliable method is developed to deal with the broken rotor bar (BRB) fault detection (FD) of a three-phase squirrel-cage induction motor (IM). The proposed method is based on an advanced IM model, which is developed using magnetically coupled multiple circuits approach. The developed squirrel-cage IM model is directly applied to the numerous computer simulations, with healthy and faulty rotor bars, in order to effectively extract the most relevant BRB feature components from the motor current and speed spectra. Thus generated discriminative BRB features are used to train an intelligent FD system based on an artificial intelligence, such as artificial neural network and support-vector machine. Finally, the method is tested and verified with BRB features obtained from additional computer simulations of the IM with healthy and faulty rotor bars. The classification results show that the proposed method can identify BRB fault with good accuracy.