| Summary: | 碩士 === 國立臺灣大學 === 化學工程學研究所 === 98 === Batch reactive distillation (BREAD) is an attractive process alternative which combines the advantages of reactive distillation and the flexibility of batch processes. There are three basic batch distillation column configurations: (1) Conventional batch distillation column (CBD): Feeds are charged at the bottom and products are taken out at the top. (2) Inverted batch distillation column (IBD): Feeds are charged at the top and products are taken out at the bottom. (3) Middle-vessel column (MVC): Feeds are charged in a middle-vessel of the column and products are withdrawn at the top and the bottom.
Tung and Yu (AlChE J., 53, 1278-1297, 2007) used an ideal system to study the effect of relative volatility ranking on the design of column configuration for continuous RD systems. A reversible reaction A+B <=> C+D is considered and constitutes a quaternary system which has 24(4!) possible rankings according to the relativity volatility among reactants and products. They further grouped these 24 possibilities into 6 (24/2!/2!) distinct categories since the two reactants and two products are interchangeable. In this work, these 6 distinct volatility rankings are applied to the 3 basic BREAD column configurations to study the effects of relative volatility ranking on BREAD process design. The process design focuses on the choice of column configuration and the optimal collection policy including when and where to collect products and off-cuts and the corresponding reflux profile. The designs are optimized based on the batch capacity (CAP) defined as the total quantity of products meeting the purity specification produced divided by the total batch time. The results indicate that if one of the reaction products is the lightest key, a CBD column can achieve the separation objective. On the other hand, if one of the reaction products is the heaviest key, an IBD column can achieve the separation objective. Moreover, a MVC with a proper collection policy shows better performances in most of the cases.
Furthermore, we investigate a new column configuration: a modified MVC in which the location of the reaction vessel is not exactly in the middle of the column. We consider the general case where the reaction vessel can be connected to the column at any point. If the reaction vessel is connected all the way at the bottom, we recover the CBD process whereas if it is connected all the way at the top, we recover the IBD. Finally, for each relative volatility ranking, we propose the most suitable column configuration and the collection policy with corresponding reflux profile which give the highest CAP.
|
|---|
