A new multivariate statistical process monitoring method using modified fast ICA

Principal component analysis is an effective multivariate statistical process monitoring approach and substantial industrial applications have been reported in recent year. This method assumed that process variables have normal distributions, which unfortunately are often invalid in real situations. A new approach based fast point independent component analysis (FastICA) is proposed without assuming that the latent variables subject to distribution. Furthermore, the number of independent components (ICs) is chosen by the sequence of non-Gaussian measure. Then we can monitor the ICs and decide the process state whether “in control” or not. The monitoring performance of the proposed method and that of the PCA-based method are compared with application to the Tenessee Eastman process (TE process). The result shows the superiority of the proposed modified FastICA (MF-ICA)-based method over the PCA-based method, and less false alarm rate can be obtained.