CC-03-157 Multivariate Process Monitoring and Early Fault Detection (MSPC) using PCA and PLS

Description:

The masses of data measured in process industry are mainly used for the purpose of process control. However, they are also useful for (a) the overview of the processes (process monitoring), (b) for fault (upset) detection, (c) for the prediction of product or process properties (soft sensors), and (e) for improved process understanding. The multivariate analysis (MVA) to accomplish these objectives has been proved as a practical and useful method. Multivariate Statistical Process Control (MSPC), process monitoring based on multivariate statistics works well both for continuous processes (e.g., distillation in a refinery, polymerization plant, or Kamyr pulp digester), and for batch processes (e.g., spray-coating, fermentation, wafer etching, cosmetic formulation).