Modeling Of Slurry Polyethylene Processes Using Aspen Plus

Developing a mathematic model can be an important method to improve industrial processes. An appropriate model can be applied to identify manufacturing bottleneck, design new grade and optimize the grade-transition strategies. In this thesis, both steady and dynamic models have been developed for an industrial slurry ethylene polymerization process. They have taken into account the complexity and characteristics of the process apparatus, polymerization system properties, and polymerization kinetics mechanism. The main results are as follows:An appropriate property model for the ethylene polymerization system was developed. The polymerization system included ethylene, hydrogen, hexane, polyethylene, catalyst and cocatalyst. Phase equilibria were calculated using the PC-SAFT state equation. Regression of the literature property data yielded the unary and binary parameters of the PC-SAFT equation of state;the reparameterized PC-SAFT equation of state was used to calculate the component properties and phase equilibria. The calculated results matched those of the literature very well. Subsequently the gas phase ethylene-to-hydrogen molar ratios of the industrial process were calculated and were found to be very close to the measured ones.A kinetic model was developed for the industrial slurry polyethylene process. It was assumed that the Ziegler-Natta catalyst was of multiple catalyst sites. The single site kinetics was acquired from the literature data on HDPE yield and ethylene conversion. Deconvolution of the gel permeation chromatography curves based on the Schulz-Flory most probable chain-length distributions yielded 5 types of active sites and the relative amount of the polymer produced by each type. The model that included a single set of kinetic and thermodynamic parameters predicted well the main properties of the polymer and main variables of the process both for grades with configurations in parallel and in series.A relationship between melt index and polymer molecular weight distribution was established. The parameters of the B-R equation and Kim 1 and Kim 2 were obtained by regression. Among those three equations, Log (MI) = 20.69-4.04log Mw -0.0044 (PDI) gave the most satisfactory results.A steady model was developed, and used for designing a new grade. The variation of the melt index was plotted as a function of the hydrogen input. The value of the hydrogen input corresponding to the desired melt index was then determined.A dynamic model was developed. The above steady model was simplified, introduced to "dynamic plus" upon adding appropriate controllers. The resulting dynamic model was able to simulate a real grade-transition process. It was used to predict transition time, production rate and the variation of the HDPE melt index as a function of load changes.