Genetic Algorithm Methodology for Broken Bar Detection in Induction Motor at Low Frequency and Load Operation

Abstract

Broken rotor bar (BRB) detection in induction motors (1M) is a challenging task because the associated failure frequencies appear near the fundamental frequency component (FFC). This identification becomes harder when the IM operates at a low frequency or with low load conditions. Therefore, techniques like motor current signature analysis may suffer on properly detecting the existence and the severity of the fault. In this sense, suppressing the FFC results helpful to improve results in the condition monitoring of IM operating at low load. This work proposes the use of a genetic algorithm for estimating and suppressing the FFC in the current signals from an IM with a BRB. Experimental results prove that the use of this technique results in better and easier identification of BRB even when the motor works at low frequency or with a low load.