(103c) Learning from Historical Data: A Key to Process Improvement and Optimization

Process and laboratory computers collect large amounts of data.  Knowing how to extract actionable information from these data is crucial. This seminar will look at recent advances and applications in the industrial use of multivariate data analysis methods for process improvement and optimization.

A brief conceptual introduction will be given to the latent variable / chemometric modeling methods used and why they are so much more suitable for the analysis of historical process data than traditional statistical methods.

Industrial examples will include the identification of important process and raw material property effects on final quality in a continuous process, and troubleshooting and optimizing a batch specialty chemical process.