Modeling and control of packed reactive distillation columns

Reactive distillation is a hybrid process that combines reaction and distillation into a single unit to significantly reduce capital cost and energy consumption, and to increase the reactant conversion. Reactive distillation has been successfully commercialized for the production of methyl tert-butyl ether (MTBE) and other chemicals. However, the strong interaction between reaction and distillation makes modeling and control of reactive distillation a very challenging task.; The main purpose of this research project is to develop steady-state and dynamic models, to validate these models with experimental data, and to develop control schemes for packed reactive distillation columns. Both the equilibrium and rate-based models were developed for steady-state simulations of reactive distillation. An intrinsic relationship between the equilibrium model and the rate-based model was found, proving mathematically that the equilibrium model is a special case of the rate-based model. Simulations of conventional distillation and reactive distillation for the tert-amyl methyl ether (TAME) and methyl acetate systems were carried out to verify the above relationship. An approach was proposed to improve the convergence of the rate-based model drastically by using the relationship between the two models. Both the equilibrium and rate-based models were validated with experimental data for the methyl acetate system, and good agreement was found between the experimental data and the simulation results.; The dynamic behavior of the reactant conversion was found to be highly nonlinear and unconventional. An approach was proposed to simplify the dynamic rate-based model drastically by assuming constant mass transfer coefficients.; The control of a packed reactive distillation column for the TAME system was investigated. Three control schemes with PI controllers were proposed. It was found that the best scheme is to control the TAME purity with a PI controller and to maximize the reactant conversion using a dynamic optimizer.; A Nonlinear Model Predictive Control (NMPC) approach was successfully implemented in Aspen Custom Model (ACM) to control the TAME purity and the reactant conversion. The performance of the NMPC controller is superior to that of the best control scheme using PI controllers.