SIMULATION OF A MULTICOMPONENT DISTILLATION COLUMN WITH CHEMICAL REACTION

Abstract

Several organic reactions are carried out in liquid phase using a homogeneous catalyst. These reactions can be easily carried out in a boiling phase using an appropriate distillation column due to which, during the course of the reaction, separation of the reaction mixture components occurs and much higher conversion can be obtained in comparison to traditional batch and continuous flow reactors. Reactive distillation combines the properties of a reactor and distillation tower in one piece of equipment. This combination of processes is suitable for systems in which chemical reactions occur at temperatures and pressures suitable to component distillation. Continual removal of products from the reaction mixture via distillation favorably alters the equilibrium, i.e. increases the conversion and minimizes undesirablu chain and side reactions. The model equations describing the behavior of reactive distillation column are nonlinear which compels the use of a variety of iterative solution methods. In this thesis, the Inside-outside tearing method. has been applied to the reactive distillation column operating at steady state. In this methoc4 two factors namely Relative volatilities and Stripping factors are defined. These factors include the effect of vapor-liquid equilibrium ratios and vapor and liquid flow rates. All the model equations are merely rearranged to give equations which may be linear or nonlinear depending upon the assumptions used in preparing the model and solved in a specific order. The production of ethyl-acetate in a 13 tray column is used as a test example. The temperatures, flow rates and concentration profiles calculated are compared with the available data. Parametric studies . for several different operating parameters are conducted to study their affect on the performance of reactive distillation column. Besides, considerable effort has been put in the modification and application of a new method called " NONLINEAR ORTHOMIN "' for solving a set of non-linear equations. This method which minimizes the square of the norm of the function vector in conjugate directions, has been recently presented and analyzed by Chronopolous (1992) and its application has been demonstrated for the solution of simple nonlinear equations. When the method was applied, using the approximations suggested, to solve six difficult problems, reported in literature as standard test problems, it was observed that this method with the approximatiohs suggested did not work for any one of them. Due to this reason, some modifications to -Non-I-i-near-Orthomin haVe been proposed in this thesis, which successfully worked to solve all the six problems. For the calculation of the step length, Chronopolous (1992) has suggested the use of inexact line search. Nonlinear Orthomin has been modified to calculate the step length based on the inexact line search of Dennis et al (1983). Another modification, based on the modified Marquardt method Edgar et al (1989) , has been made to the Nonlinear Orthomin method, to force the Hessian matrix to be positive definite. It is our view hat this method may be used in future to solve non-linear equations, which frequently arise in the modeling of distillation columns with or without chemical reaction. vii