A Dynamic Response Surface Model for Polymer Grade Transitions in Industrial Plants

The dynamic response surface methodology (DRSM) has recently been applied to accurately model the dynamics of polymer grade transitions (Wang and Georgakis, Ind. Eng. Chem. Res. 2017, 56, 10770-10782). Following this publication, we propose here two methodological improvements that greatly reduce the required number of experiments and still achieve similar model accuracy. First, a novel design of the experiments is defined to combine the previously separate input domains incorporating transitions that increase and decrease a measure of polymer grade. This enables use of a single DRSM model, instead of the prior two separate ones, and reduces the number of experiments by half. Furthermore, a sequential modeling strategy is proposed, appending experiments to an initial simpler design to estimate a more complex DRSM model. As the desired DRSM model complexity is not known a priori, the sequential modeling strategy is critical in achieving the desired model accuracy with the minimum number of experiments. These methodological advances have been tested against an academic and an industrial process simulation.