A Novel Experimental Method to Detect Early Rotor Faults in Induction Machines

Abstract

The Induction Machines (IMs) have numerous industrial interests, as they have an uncomplicated design, strong, robust, and vulnerable to maintain. Despite that fact, a small fault in one rotor bar or a low Unbalance Rotor (UR) causes the break of this bar. Subsequently, the break of the adjacent bar because of the mechanical vibration generated by this imbalance of the rotor. This contribution proposes a new method based on the analysis of the stator line current to enhance the reliability of the monitoring system of an incipient defect in one rotor bar of the Induction Motor (IM). This new method, which combines the Electrical-Time-Synchronous-Averaging (ETSA) procedure with the Vector Park Transform (VPT) technique and Motor Current Signature Analysis (MCSA), and a Fuzzy Logic Algorithm (FLA), is performed. Indeed, an incipient rotor failure cannot be detected using the stator line current signal curve. Absolutely, in cases of the minor defect in one only rotor bar and low UR at low load. Nevertheless, the residual energy contained in each spectrum is considered as input variables for a fuzzy logic system to classify the type and the level of rotor defects. The proposed algorithm applied to the measured stator line-current has been built under the MatLab environment. The experimental obtained results confirm that the conceived algorithm can classify exactly the rotor faults.