Research on Detection Method and Test of Generator Imbalance Fault

Based on techniques of time-domain index feature extraction and principal component analysis (PCA), a method for unbalanced fault diagnosis is proposed in this paper. Five dimensionless factors corresponding to IMF are calculated after the original data is decomposed (EMD) by empirical mode, then the PCA is applied to the data vectors, which are processed by dimensionality reduction and residual space projection. At last, fault diagnosis test of balance loosening was carried out on the test bench of intelligent water supply system. Test results show that, the established PCA model is effective in judging unbalanced faults.