Stator Inter-turns Short Circuit Fault Detection in DFIG Using Empirical Mode Decomposition Method on Leakage Flux

Abstract

Inter-turn short-circuit fault (ITSCF) in doubly fed induction generator (DFIG) effect the performance of the wind turbine, so the surveillance of DFIGs is of very importance for the safe operation of the wind turbine. To solve this issue, an experimental setup was constructed for the detection of stator ITSCF. This experimental setup consist of three-phase wound rotor induction machine (WRIM), monitoring systems, frequency converters and related control and a drive motors. This experimental system can truly simulate the actual working conditions of the machine when the machine is short-circuited, and the experimental simulation can be closer to the operation process of the double-fed induction generator. Flux monitoring method was used to detect the ITSCF in stator windings. Time domain analysis was used on flux signal to detect the ITSCF at its initial stage of the fault. Empirical Mode Decomposition (EMD) method was applied on flux to detect the stator ITSCF. The EMD of the flux signal in the stator winding ITSCF yields 9 layers of empirical modal components IMFs. The IMF5 and IMF6, was found useful to discriminate the normal and faulty condition of the machine. The results show that EMD analysis on flux monitoring is advance and effective way than the other methods to detect the ITSCF in stator windings.