Data-based PI Control Strategy of a Polymerization Reactor

C. Cheng and M.-S. Chiu (Singapore)

Keywords

Justintime learning, PI, Polymerization reactor

Abstract

Modeling and control of polymerization reactors present challenging tasks in process control practice. In this paper, a nonlinear PI design strategy derived from just-in time modeling framework is developed as an attractive alternative for polymerization reactor control. The proposed data-based PI design is tested by using a literature example of a polymerization reactor, where an isothermal free-radical polymerization of methyl methacrylate is carried out using azo-bis-isobutyronitrile as initiator and toluene as solvent and a comparison with a Volterra-model based predictive control is made.

Important Links:



Go Back