The thesis starts with a brief overview of unsaturated polyesters. In particular, the usage of raw materials, the application of unsaturated polyester resins, and, the worldwide supply and demand of the unsaturated polyester resins are discussed. Unsaturated polyester is traditionally produced in a batch-wise-operating reaction vessel connected to a distillation unit. The total production time is around 12 hours and often leads to batch-to-batch inconsistency. Process intensification is required for the unsaturated polyester process to reduce the production time and to achieve a better quality of the product. An attractive alternative to batch-wise polyester production is reactive distillation. In chapter 1, the attractiveness of reactive distillation for the synthesis of unsaturated polyester is discussed. The goal of the thesis is to develop and evaluate a reactive distillation process for the production of unsaturated polyester from anhydrides and glycols. To accurately predict the behavior of reactive distillation process, reliable kinetic and thermodynamic models are required. Therefore, in chapter 2 a dynamic model for a batch-wise operating reaction vessel connected to a flash separation unit is developed in order to validate the kinetic and thermodynamic models and their parameters. This model includes kinetics, description of the change of rate order during the reaction, the polymer NRTL non-ideal thermodynamic model based on non-random theory of liquid (NRTL) and mass balances. The reaction between maleic anhydride and propylene glycol has been taken as a case study. The reaction scheme is complex and the proposed model takes four types of reactions into account; ring opening, polyesterfication, isomerization and saturation reactions. The acid value of the polyester, number-average molecular weight, distilled mass and glycol concentration in the distillate have been subsequently used to validate the model and the model predicts these important variables reliably. The process description is improved by using the vapor liquid equilibrium data predicted from the polymer NRTL model. After successful validation of the kinetic and thermodynamic models, the feasibility of the reactive distillation process for the unsaturated polyester is presented in chapter 3. Moreover, the simulation results of reactive distillation model are compared with the batch reactor model simulation results to determine advantages gained by the reactive distillation over the traditional batch process. The simulation study shows that the total production time of polyester in a continuous reactive distillation system is reduced to 1.8-2 hours compared to the12 hours of the industrial batch reactor process. The model demonstrated that reactive distillation has the potential to intensify the process by factor of 6 to 8 in comparison to the batch reactor process. After finding that reactive distillation is an attractive alternative for the polyesters synthesis, a more in depth analysis is performed. Particularly, the influence of the liquid back mixing on the description of the reactive distillation process, product transition time, the amount of undesired product formation during the product changeover is investigated. Since the current state of the art modelling approach does not account for liquid back mixing, the rate-based model is extended to account for liquid back mixing. The simulation results of extended rate-based model demonstrated that axial dispersion significantly influences the reactive distillation process and cannot be neglected. On the basis of current research work and literature review, a novel design methodology for the economical and technical evaluation of reactive distillation is proposed in chapter 4. Moreover, the applicability of various design methods for reactive distillation is discussed. The proposed framework for the economical evaluation determines the boundary conditions (e.g. relative volatilities, target purities, equilibrium conversion and equipment restriction), checks the integrated process constrains, evaluates economical feasibility, and provides guidelines to any potential reactive distillation process application. Providing that a reactive distillation process is economically attractive, a technical evaluation is performed afterward in order to determine the technical feasibility, the process limitations, working regime and requirements for internals as well as the models needed for reactive distillation. This approach is based on dimensionless numbers such as Damkohler and Hatta numbers, as well as the kinetic, thermodynamic and mass transfer limits. The proposed framework for economical and technical evaluation of reactive distillation allows a quick and easy feasibility analysis for a wide range of chemical processes. Several industrial relevant case studies (synthesis of di-methyl carbonate (DMC), methyl acetate hydrolysis, toluene hydro-dealkylation (HDA) process, fatty acid methyl esters (FAME) process and unsaturated polyesters synthesis) are used to illustrate the validity of the proposed framework. In chapter 4, it is found that the bubble column is the potential device for producing unsaturated polyesters by the reactive distillation. Moreover, the introduction of packing or partition trays in the bubble column significantly improves the unsaturated polyester process because packing or partition trays provide a better mass transfer and the multi-stage effect in the column. But considering the lack of information about the behavior of counter-currently operated bubble columns in the presence of structured packing or partition trays and in a viscous system, a systematic investigation on the gas holdup, axial dispersion and mass transfer in the packed bubble column and the trayed bubble column is undertaken in chapter 5. Four different types of structured packings (Super-Pak, Flexipac, Mellapak and Gauze) and two types of perforated partition trays (with 25% and 40% tray open area) are used to characterize the packed and trayed bubble column, respectively. It is observed that the packed and trayed bubble columns improve the gas holdup and mass transfer compared to the empty bubble column and reduces the axial dispersion significantly. Particularly, the Gauze packing improves the gas holdup and mass transfer and, sufficiently reduces the axial dispersion. In contrast, Super-Pak offers only a modest improvement because of its open structure. Comparison of the experimental data of the packed and trayed bubble column indicates that the partition trays improve the bubble column in the same order as packing. The gas holdup, axial dispersion and mass transfer depend more strongly on the gas velocity compared to the liquid velocity. The liquid viscosity also significantly influences these parameters and therefore the empirical correlations obtained from the air-water system cannot be applied for the viscous system. Moreover, experimental data of the packed, trayed and empty bubble column are correlated by dimensionless numbers. Empirical correlations for the gas holdup, Bodenstein number (for the axial dispersion coefficient) and Stanton number (for the volumetric mass transfer coefficient) as a function of the Froude and Gallilei dimensionless numbers are proposed. In chapter 6, an experimental pilot plant validation of the reactive distillation process for the polyester synthesis is presented. Two different configurations are investigated: 1) a reactive distillation column and 2) a reactive distillation column coupled with a pre-reactor. Due to a relatively short residence time of 0.32 hours and an operating temperature of 190oC in case of the first configuration, a maximum conversion of 37% was achieved; which indicates monoester formation in the reactive distillation column. In the case of the second configuration, a 90% conversion is achieved within 0.55 hours at a temperature of 250oC in the reactive distillation column coupled with a pre-reactor; which confirms the polyester formation in the reactive distillation column. The extended rate-based model developed in chapter 3 is used to simulate the pilot reactive distillation column. The model predicted the experimental data (acid value, conversion, isomerization and saturation fraction, number-average molecular weight, the degree of polymerization and water fraction in the distillate) adequately (5-22%). Moreover, the product specifications of the polyester produced at 250oC in the reactive distillation column is in the range of polyesters produced in the traditional industrial batch reactor setup. Furthermore, discoloration of the polyester was hardly noticed even though the column was operated at 250oC. Finally in chapter 7, the validated model is used to find the best suitable internal and feed configurations of the reactive distillation process for unsaturated polyester synthesis. Moreover, multi-product simulations are performed to find the operational parameters for producing two different grades of polyester in the same equipment. Finally, the product transition time during product changeover is determined. The criteria to select the best configuration are minimum volume and energy requirement to produce 100 ktonnes/year polyester. First the best suitable internal for the column is identified and then the best suitable feed configuration is identified. From simulations, we concluded that the configuration which contains the reactive stripping section as a packed bubble column and the reactive rectifying section as a packed column requires minimum volume and energy to produce 100 ktonnes/year polyester. With respect to the feed configuration, we concluded that the feeding of monoesters to the reactive distillation column significantly intensifies the polyester process compared to an anhydrous reactant fed to the column. Moreover, the product transition time in this configuration is also significantly lower compared to the other configurations. In conclusion, a reactive distillation column coupled with a pre-reactor is the most promising alternative to continuously produce unsaturated polyesters. It requires a factor 10 (90%) lower volume, a factor 15 (93%) lower production time and a factor 3 (66%) lower energy as compared to the traditional batch reactor process to produce 100 ktonnes/year of polyester. Hence, the reactive distillation process improves the unsaturated polyester synthesis in all domains of structure, energy and time compared to the traditional batch reactor process coupled with a distillation column.
|Kwalificatie||Doctor in de Filosofie|
|Datum van toekenning||9 nov 2011|
|Plaats van publicatie||Eindhoven|
|Status||Gepubliceerd - 2011|