Modeling of the Industrial Spherizone Process for Impact Copolymer Polypropylene: Forward and Reverse Engineering Applications

The emerging Spherizone technology employs three reactors operating in series for impact copolymer polypropylene (ICP) production, and this complex process configuration makes tailoring of ICP properties challenging. To address this issue, a bidirectional relationship between the polymerization conditions and molecular properties is established in this study. The proposed bidirectional model couples a kinetic-based moment method, three-stage cascade reactor models, polymer property distribution models, and a Bayesian optimization algorithm. In particular, industrial molar mass distribution (MMD) data are used to deconvolute the number of catalyst active sites, while average polymer properties are employed to estimate the kinetic parameters. The simulation results reveal that key parameters in each reactor of the cascade system exert significant influence on the polymer properties. The proposed model is capable of inversely predicting the polymerization conditions in each reactor required to achieve the desired MMD and demonstrates its predictive capability for the inverse design of MMD targets that have not been previously simulated. The modeling framework can be further extended to other polymerization systems.