Detection of Interturn Short-Circuit Faults in DFIGs Based on External Leakage Flux Sensing and the VMD-RCMDE Analytical Method.
Shouwang Zhao,Yu Chen,Attiq Ur Rehman,Feng Liang,Shuang Wang,Yong Zhao,Wei Deng,Yong Ma,Yonghong Cheng
DOI: https://doi.org/10.1109/tim.2022.3186061
IF: 5.6
2022-01-01
IEEE Transactions on Instrumentation and Measurement
Abstract:Fault detection based on external leakage flux can address various faults, such as short-circuit faults in stators and rotors. External leakage flux sensing technology, as a noninvasive method, has been attracting increasing attention and research. However, for a high-power doubly fed induction generator (DFIG), the external leakage flux signals of the generator are easily overwhelmed by a strong noise background, and the fault diagnosis and recognition of stator interturn short circuit (S-ITSC) and rotor ITSC (R-ITSC) are complicated, which limit the practical engineering application of leakage flux sensing technology. Aiming at addressing ITSC faults in DFIGs, this article proposes a method for fault feature extraction and recognition of stator and rotor short circuits based on variational mode decomposition (VMD) and the refined composite multiscale dispersion entropy (RCMDE) analytical method for the external flux leakage of the generator. The optimized parameters of VMD are selected automatically by the genetic algorithm (GA), and VMD is adopted to adaptively decompose the external leakage flux signals into a series of intrinsic mode function (IMF) components. The evaluation criterion based on the correlation number in the frequency domain combined with the autocorrelation function is used to select the best IMF components with clear features. Two different components are chosen to reconstruct the flux leakage signal. The characteristic frequency of the reconstructed signal is analyzed by the Hilbert-Huang transform (HHT) and the total harmonic effective value of the characteristic component. To effectively identify and diagnose the generator S-ITSC and R-ITSC faults, RCMDE is used for the flux leakage reconstructed signal. The experimental results under different stator and rotor short-circuit levels show that this diagnosis method can effectively extract the weak feature information from the external flux leakage signals and perform fault feature extraction and recognition for stator and rotor short circuits.