Model reduction and optimization of a reactive dividing wall batch distillation column inspired by response surface methodology and differential evolution

Abstract

Carrying out reaction and separation simultaneously in a reactive dividing wall batch distillation column batch RDWC in the case of ethyl acetate synthesis provides the possibility of separating both products and increasing the equilibrium reaction conversion. Overcoming the known azeotrope conditions, high purity for ethyl acetate and decreasing the batch time compared to simple reactive batch distillation are the advantages of this configuration. The corresponding dynamic simulation is carried out by simultaneously solving the model-associated system of differential and algebraic equations. In this study, the optimal values of the vapour and liquid split ratios are considered as the decision variables in order to maximize the amount of ethyl acetate accumulated during batch time. The optimization strategy is implemented inspired by response surface methodology in which an optimal surface is fitted to the collected data set using differential evolution (DE). The optimal surface relevant algebraic equation is then considered as the reduced form of the complex model and the optimal values are obtained using the DE method.