Abstract
•A generalized likelihood ratio test (GLRT) based online reduced kernel PLS (KPLS), named OR-GLRT, is developed.•OR-GLRT fault detection approach is proposed to enhance chemical process monitoring.•The detection performance of the new chart is studied using Tennessee Eastman process (TEP).•The detection results are evaluated using three fault detection criteria: the missed detection rate(MDR), the false alarm rate (FAR) and the computation time (CT).
In this paper, an improved fault detection method is proposed based on kernel partial least squares (KPLS) model and generalized likelihood ratio test (GLRT) detection chart in order to enhance the monitoring abilities of nonlinear chemical processes. To deal with both high computational cost for large data set and time-varying dynamics of industrial processes, the proposed method is used to select a reduced set of kernel functions to build the KPLS model and applies it for online fault detection based on GLRT chart. Comparing with the conventional KPLS technique, the proposed method has the advantages of improving the computation efficiency by decreasing the dimension of kernel matrix. The developed online reduced KPLS-based GLRT (OR-GLRT) could improve the fault detection efficiency since it is able to track the time-varying characteristics of the processes. The fault detection performance of the developed OR-GLRT method is evaluated using a Tennessee Eastman process. The simulation results show that the proposed method outperforms the conventional GLRT technique.