Application of Hybrid Modeling in Polymer Processing
Sprache des Titels:
ANTEC 2020 The Virtual Edition
or many decades, the setup and solution of polymer processing models involved use of analytical or numerical methods. These characteristics have changed with the recent digitization of polymer processes and the collection of enormous amounts of data. It is increasingly common to use data-driven modeling techniques to analyze processes, for which analytical and numerical models may not fully describe the process behavior in operational situations. These techniques have significantly extended the set of tools available to the engineer, providing new possibilities of how to develop more accurate process models. As a result, the setup of an appropriate modeling strategy more than ever requires a thorough understanding of the individual modeling techniques. This article was designed to address the potentials and limits of analytical, numerical, and data-based modeling techniques when modeling polymer processes. Moreover, we show how these methods can be combined into one hybrid approach to solve polymer process models not solvable so far. The findings are further illustrated by means of a particular use case, which models the flow of polymer melts in single-screw extruders.