The Integration of the Synthesis Methodology in the Design of a Five-Component Distillation Sequence

Abstract

The separation of a multicomponent stream by distillation represents a common problem in the industrial field. The possibility to realize the same separation task by means of different sequences have opened an unique research topic focused on the definition of a complete searching space that includes all the possible alternatives. The searching space can be generated following different approaches based on a mathematical background, like superstructures, or based on the connection of the simple column sequences with the alternatives generated. The correspondence between simple column sequences and the alternatives predicted is the principle of the generation method followed in the present work. Using this type of approach the generation and the design method used to screen and evaluate the alternative’s performance are deeply related. The correspondence between the functionality of the column’s sections of the simple columns and the corresponding alternatives was used to get the initialization parameters, then these parameters were optimized by a sensitivity analysis. The proposed design method was called Sequential Design Method (SDM). The method was applied for a five component separation sequence considering a particular class of alternative configurations called Modified Simple Column (MSC) sequences. In order to test the proposed methodology, the same sequences were designed using an optimization procedure based on a hybrid multi-objective (HMO) algorithm. It was found that the SDM is a fast and reliable design procedure that allows the user to easily define the column sequences‘ parameters. Moreover the SDM can be used to correctly initialize the multi-objective algorithm. When the HMO algorithm is initialized with the design parameters obtained from the SDM, the computational time is reduced by 21 % compare to a random initialization.