Fault isolation by partial pca in the nonlinear process

Principal component analysis(PCA) is a popular tool in fault detecting of the complex plant, but offers little support on fault isolation. Partial PCAPPCA) well improves fault isolation capability by utilizing structured residuals' .isolation property. To overcome the nonlinearity limitation in the process industry, PNPCA based on nonlinear PCA and PPCA is proposed and the implementing steps are demonstrated. The simulation on Tennessee Eastman (TE) process shows the attainability of this method.