Novel kinetic modeling strategy for industrial-scale UNIPOL polypropylene with Ziegler-Natta catalyst

Polypropylene (PP) is a synthetic resin with exceptional physical and chemical properties, making it suitable for various household and industrial applications. However, as the demand for PP continues to rise, the polypropylene industry faces challenges in improving its quality and efficiency. Developing an accurate mathematical model that describes the continuous polymerization process is crucial for enhancing PP production, as it is useful for controlling product quality and improving production efficiency. While previous studies have focused on modeling propylene polymerization using laboratory-scale and pilot-scale reactors, these models do not effectively apply to industrial-scale processes due to the complexity of the industrial environment. Therefore, a novel kinetic modeling method combining the modeling of the kinetic mechanism of the polymerization reaction with an intelligent optimization algorithm for the industrial-scale continuous gas-phase fluidized bed propylene polymerization process is proposed. To establish the kinetic mechanism model, the improved moments (MOM) method is utilized to derive a set of mass balance equations that account for various fundamental processes such as catalyst activation, chain initiation, chain growth, chain transfer to hydrogen, chain transfer to monomers, chain transfer to co-catalysts, and spontaneous deactivation. However, kinetic parameters serve as the link between the theoretical model and the actual production process, so the kinetic parameters estimation problem need to be transformed into a multi-objective problem by minimizing the least squares difference between the measurement data of the actual industry process and the predicted result of the kinetic model. Therefore, an improved multi-objective mayfly algorithm (IMOMA) is introduced to estimate the polymerization kinetic parameters, which incorporates the simulated binary crossover (SBX) operator and the polynomial mutation (PM) operator into the basic multi-objective mayfly algorithm (MOMA). Finally, by fitting industrial measurement data of the melt index (MI) and the polymerization rate (PR) of an actual polypropylene industrial case study, the proposed algorithm accurately determines the polymerization kinetic parameters. The results demonstrate that the MOM-IMOMA modeling approach outperforms other methods and holds significant promise for industrialscale polypropylene production.