Multivariate Statistical Process Control of Batch Processes Using PCA and PLS

Multivariate statistical procedures for monitoring the progress of batch processes are developed. The only information needed to exploit the procedures is a historical database of past successful batches. Multi-way Projection to Latent Structures is used to extract the information in the batch set-up data and in the multivariate trajectory data, by projecting them onto low dimensional spaces defined by the latent variables or principal …