Fault Detection and Isolation Based on MM- ICA with Application to High-speed Railway

摘要

In this paper, a new fault detection and isolation scheme based on multiple models independent component analysis (MM-ICA) is established for high-speed railway. The proposed method firstly uses independent component analysis (ICA) to extract the essential independent components (ICs). Then a novel idea, which is to construct multiple ICA models, is proposed to locate fault. The proposed method is applied to fault detection and identification in both a simple multivariate system and highspeed railway motor, and the results are compared with principal components analysis (PCA). The simulation results clearly show the power and advantages of the novel fault detection and identification method based on MM-ICA.