Research On Process Monitoring For Complex Industrial Processes Based On Multivariate Statistical Analysis

Automatic control systems for modern industrial processes have been expanding to complexity and larger scale. However, the complexity of system will result in unexpected faults which will bring large economical loss or even cause human injuries and environmental problems. Therefore, research on integrated process control system, which includes such roles as control, monitoring, diagnosis, proves to have important theoretical and practical value.Process monitoring based on multivariate statistical analysis is an important branch in the research of process control system. In this dissertation, important aspects of process monitoring based on multivariate statistical analysis are presented and studied systematically. The main contents of this dissertation are described as follows:In this paper, process monitoring methods are summarized. Then, the mathematic tools applied to process monitoring based on multivariate statistical analysis, including PCA, KPCA, ICA and KICA, are briefly introduced. Then, their applications to process monitoring are introduced.In this paper, the correlation between variables are studied before process monitoring. Multivariate regression analysis is used to judge the correlation between gaussian variables. Then, process monitoring models based on PCA and KPCA are established, by taking the TE benchmark as the application background. The simulation results indicate that the proposed method is feasible. The ICA-based method is used to judge the correlation between non-gaussian variables. Then, process monitoring models based on ICA and KICA are established, by taking the pernecillin fermentation process as the application background. The simulation results indicate that the proposed method is effective.S.W.Choi et al propose a fault identification method, which combines the KPCA-based denoising and the PCA-based fault identification. In view of the shortage of the method proposed by S.W.Choi et al, an improved scheme for KPCA-based fault identification is presented in this paper. Then, simulations are carried out for the improvd method. The simulation results indicate that this method not only can identify the fault caused by single variable, but also can identify the fault caused by several variables simultaneously. And the improved method has reduced the operations.