(136d) Development of a Dynamic Model for the Production of Impact Modified Styrene/Maleic Anhydride Copolymers

The impact modified Styrene/Maleic Anhydride copolymer is produced using one stirred tank reactor system where feed streams consisting of styrene, maleic anhydride and dissolved rubber are continuously mixed. This system was simulated using the Aspen Custom Modeler (ACM) software package together with the Polymer Plus software to provide the polymer properties. The dynamic model features a Continuous Stirred Tank Reactor (CSTR) and the polymerization is based on the first principles with a kinetic scheme of free radical polymerization. The kinetic parameters were estimated based on experimental data from the pilot plant. The Polymer Non-Random Two-Liquid (Poly-NRTL) activity coefficients models were used to provide the thermodynamic properties (fugacity coefficient, enthalpy, entropy and free energy). The Van Krevelen thermo-physical property models were used to calculate density, enthalpy, entropy, Gibbs free energy and heat capacity for the polymers and segments. One of the objectives of this modeling effort was to create a tool to help predict reactor performance during unsteady state operating conditions and to address reactor performance issues such as, effect of varying the reactor temperature and the feed composition on the conversion and polymer properties. Thus, optimizing and improving the product value through better reactor control. Input to the model includes the feed composition and reactor operating conditions. Output from the model includes: polymer production rate, polymer composition, and polymer molecular weights (Mn & Mw). Model predictions for conversions (percentage solids) and the polymer compositions were within ±10 % compared to the experimental data. For Mn, model predictions in 73% of the cases were within ±10 % of the experimental data and in 85% of the cases were within ±15 % of the experimental data. For Mw, model predictions in 82% of the cases were within ±10 % of the experimental data and in 93% of the cases were within ±15 % of the experimental data. These results are in line with the experimental data considering the accuracy of the experimental measurements of Mn and Mw.