Information extraction from the Evolving Factor Analysis

The goal of our project (HES-SO n°11246) is to develop chemometric methods that enhance the development of chemical reactions. Semi-batch experiments are performed using a fully automated chemical reactor. Real-time analytical data, infrared (IR) spectra for instance, are syn-chronized with standard process variables [1]. In parallel to the on-line measurements, the reaction is followed by conventional chemical analysis such as gas chromatography (GC). We have chosen an aldol condensation as an example reaction. IR spectra are preprocessed (wave number region selection, filtering) and processed by Evolving Factor Analysis (EFA) [2] to monitor the chemical reaction. We will present our comparative results based on the real analysis, includ-ing the influence pretreatment, of collected data from the synthesis and on a dynamic simulation of the reaction [1] Matthias Otto in ‘Chemometrics, Statistics and Computer Application in Analytical Chemistry’, Wiley-VCH, 1999, New York. [2] Malinowski, E. R. in ‘Factor Analysis in Chemistry’, John Wiley & Sons, 1991 (2nd edition), New York.