Simulation And Optimization Of An Industrial HDPE Slurry Process Based On Equation Oriented Models

The slurry polymerization process is the most widely used method in high density polyethylene (HDPE) production, due to the mild reaction conditions, controllable operation conditions and good product quality. The HDPE slurry process contains four major aspects, thermodynamic properties of polyethylene system, modeling of polymerization mechanism, molecular weight distribution (MWD), and process simulation/optimization. Perturbed-chain statistical associating fluid theory (PC-SAFT) is the most accepted thermodynamic model. However, PC-SAFT is a complicated model and have to solved using embed sequential algorithm. It is thus difficult to be used in equation-oriented approach. Based on the polymerization mechanism, population balances of different active chain must be considered when establish the reaction model, but the rapid expansion of the model could lead to an unsolvable problem. MWD is the main factor affecting the application of products, hence the most important and microcosmic quality index of HDPE. The whole model of ethylene polymerization process will be a large-scale model with heavily coupled variables and strong nonlinearity. To solve the model efficiently and accurately becomes the major issue. Equation-oriented approach (EO) has gained rich research interests in simulation and optimization of chemical process owing to the advantages of ease switching between simulation and optimization models, high efficiency and good convergence properties. Based on the EO approach, this thesis focuses on four aspects of HDPE slurry process, and some studies on modeling, simulation and optimization of HDPE slurry process are carried out. The contributions of this thesis include the following four aspects.1. Surrogate models based on Kriging technique are proposed and developed to describe the relationship between the physical conditions and the material properties for the PC-SAFT equations. A database of material properties has been established in the actual operating range for the second modeling of PC-SAFT. The good simulation results of several steady-states show the robustness of the Kriging models.2. Thorough studies of polymerization reaction mechanism are carried out, population balance equations and moment model equations of ethylene polymerization with the five-sites-type Ziegler-Natta catalyst are derived based on the Generating Function method. Flory distribution is approved to be justified for steady-state MWD calculation, the MWD formulations of first and second reactor are established based on Flory distribution. Also other models of auxiliary modules are established.3. Some operating conditions on a single reactor with and without reflux are simulated using the EO approach. Previous problem that Aspen Polymer Plus unable to converge in single reactor with reflux cases has been solved, which allows the steady MWD optimization. Process with twin reactors in series is also validated in this thesis. The good simulation results show the accuracy of the kinetic models, Kriging models and MWD models.4. The relationship between the MWD and kinetic parameters is analyzed; the operation condition of HDPE process is validated by sensitivity analysis. Three optimization problems for MWD are proposed, two of them are implemented in this thesis, and the good results provide a potential use in HDPE product design.