Title

Process chemometrics for dynamic systems

Abstract

Estimating (anticipated and unexpected) dynamics is a key task in process monitoring and control, and process chemometrics plays a significant role in this engineering task. But where existing modeling strategies are primarily designed for pure batch or continues production systems, most operations in the food and biotech industry run in a semi-continuous mode, including (more or less) clear transitions. In this paper both implicit and explicit use of the process characteristics will be evaluated combining “soft” model-builder and “hard” engineering ideas.

Bio

Frans van den Berg is associate professor in the Chemometrics & Analytical Technology section at the Department of Food Science, University of Copenhagen, Denmark (KU.FOOD.CAT). He is educated as a lab technician, followed by a civil engineering degree in Laboratory Information and Automation, an MSc in Analytical Chemistry and a PhD in Process Analytics and Chemometrics, and is presently trying to combine all these disciplines into one area: Process Analytical Chemistry and Technology (PACT). His main interests are in process spectroscopy and data analysis, specifically the application of chemometrics, statistics and mathematics in (process) data collection, integration, control and optimization.

“I have been active in Chemometrics since 1990, and my favourite methods are … those that fit the investigative question, or to misquote Occam’s Razor “Among competing methods that (typically) all work equally well, the one which matches most natural with the problem at hand should be selected”. Multivariate (subspace) exploratory and regression methods have been absorbed into almost all scientific areas, but I think the future of chemometrics as a separate discipline will be bright if the community keeps integrating the principles from other disciplines, much like its roots being analytical chemistry. The variance-based concepts of statistics, the time-based nature of process signals or the bio-based understanding in metabolomics studies should be an integrated part of pre-processing, modelling and post-processing data in the most optimal way. I am looking forward to be inspired by many innovative ideas and applications as SSC16!”