Kernel principal component analysis and its application to fault feature extraction

An approach to fault feature extraction is presented, which is based on kernel principal component analysis (KPCA). In this approach, the integral operator kernel functions is used to realize the nonlinear map from the raw feature space to the high dimensional feature space. By performing PCA on the high dimensional feature sets, the nonlinear principal components of the raw feature space are obtained. In succession, the selected nonlinear principal components are used to construct the feature subspace. The fault data sets of rotator test-bed are used to test the KPCA based method. The results indicate that the method is more suitable for nonlinear feature extraction from fault signals, the extracted features based on KPCA perform better fault recognition ability and they are robust for various classifiers.