Set-Membership Parameter Estimation: Improved Understanding of Microgel Polymerization

Functional microgels based on N-Isopropylacrylamide (NIPAM) are in the focus of research because their properties are promising for diverse applications, such as drug delivery and tunable membranes. However, the synthesis of microgels with optimal properties for the respective applications is challenging and time-intensive due to the sensitivity of the product properties to the synthesis conditions. Optimal synthesis strategies can be determined and controlled with model-based approaches. This in turn requires process models with parameters estimated based on experimental measurements. However, available measurement are not always sufficient to determine all model parameters. Herein, we use inline Raman and calorimetric data. We use a set-membership approach for parameter estimation and identifiability analysis. We solve a series of constrained dynamic optimization problems to approximate the feasible parameter set by a box. The results show that, while kinetic parameters corresponding to the main monomer are identifiable, a subset of parameters corresponding to the cross-linker have wide confidence ranges. This is caused by the low initial concentration of cross-linker in comparison to the concentration of monomer. The determined parameters enable the improved model-based prediction of the properties of microgels based on NIPAM and pave the way towards model-based control applications. (C) 2019, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.