Monitoring and control of batch processes

Several application studies have indicated that multivariate statistical technology can be applied batch systems to help process operators and engineers identify abnormal operation and improve product consistency through automatic control. Multivariate statistics relies significantly on the statistical routines, such as principal component analysis (PCA) and partial least squares (PLS) to perform the required processes. PCA is a technique designed to extract major features within a data set and it can identify several statistics that provide significant information regarding process plant operation. PCA involves the use of the square prediction error (SPE) method to perform monitoring process of a process plant. SPE establishes the relationships between the process variables, comparing them with those identified under normal operating conditions.