Developing a digital twin of a polymerization reaction for process optimization

The application performance of polymers is heavily influenced by their molecular weight distribution. As the dynamic modeling of the MWD in polymerization reactions is difficult based on first principles, a semi-physical or grey-box modeling approach is proposed for step-growth polymerization processes, utilizing molecular weight distribution measurement data. The method is tested on data from an industrial polymerization reactor.