Fault Detection for Process Monitoring Using Improved Kernel Principal Component Analysis

In order to detect abnormal events of chemical processes, a new fault detection method based on kernel principal component analysis (KPCA) is described. Firstly, it removes the noise from data set using wavelet packet transform (WPT). Secondly, a feature vector selector schemes (FVS) based on a geometrical consideration is given to reduce the computation complexity of KPCA when the number of the samples becomes large. Then, the denoised data is disposed using KPCA and and SPE are constructed in the feature space. KPCA was applied to fault detection. To demonstrate the performance, the proposed method is applied to the Tennessee Eastman process. The simulation results show that the proposed method effectively and quickly detect various types of faults.