Evaluation Method of Performance Degradation based on End Magnetic Flux Leakage for Short Circuit Fault of Doubly Fed Induction Generator

The health status of the Doubly Fed Induction Generator (DFIG) is related to the actual operating conditions, the external environment, the accumulation of sudden factors, and the coupling effect. The degradation feature extraction is mainly based on a single signal or multiple statistics of a single signal. The principal component components were extracted from the end magnetic flux leakage (MFL) monitoring data of DFIG to evaluate performance degradation. This paper proposes an evaluation method of performance degradation for short circuit faults based on Variational Mode Decomposition (VMD) and Support Vector Data Description (SVDD). The process for detecting short circuit faults and performance degradation by monitoring the end-external MFL. For short circuits, when the magnetic flux leakage signal monitoring by the external environment, abnormal signal and noise problems, the VMD method is used to decompose the MFL signals and extract the most relevant modal composition of Root Mean Square (RMS), Singular Value Decomposition (SVD), Sample Entropy (SE), Refined Composite Multiscale Dispersion Entropy (RCMDE) as the main feature vectors. For a set of feature vector sets of MFL signal, then using the SVDD to perform performance degradation assessment. The distance between the sample data to be checked and the center of the trained hypersphere model is used to describe the performance degradation degree, and the membership function is used to transform the distance index into the membership degree of the normal state as the performance degradation index, to realize the state evaluation of the performance degradation degree of the generator.