Fault Detection and Diagnosis Based on Principal Component Analysis

Fault detection and diagnosis method based on the principal components analysis (PCA) is discussed . The fault detection and diagnosis simulation to a typical chemical process is performed by means of statistical methods like Holleting T 2 and Q. The contribution charts are used to undertake fault diagnosis. The simulation results show that PCA is an effective approach to fault detection and can only work for the simple sensor fault diagnosis. A novel idea to combine PCA with causal model based approach is presented for the future research aiming at complex sensor fault and internal process fault.