Modeling and validation of multi-objective optimization for mixed xylene hybrid distillation/crystallization process

Abstract

Innovative hybrid distillation/crystallization processes offer significant energy saving and cost reduction for the separation of isomer mixtures. Despite of the enormous potential of hybrid separation processes, they are not widely exploited in industrial applications due to the complexity of the design and multi-objective optimization of these highly integrated processes. In this study, the mixed xylene hybrid scheme was first simulated to provide the initial value of non-dominated sorted genetic algorithm-III. The operating parameters were then optimized through the Python-Aspen platform. Coupled with the online monitoring tool, a rigorous melt crystallization experiment was developed to track crystal size distribution (CSD) and crystal shape. By comparing the optimized results with the original ones, the total annual cost and CO2 emissions were, respectively, reduced by 5.35 % and 12.80 %. Through the experimental validation, the accuracy of the simulation-based crystallization CSD prediction method was discussed, and the potential and challenges of this novel method were proposed. This work aims to minimize the total annual cost and CO2 emissions of the mixed xylene hybrid process simultaneously and explore the feasibility of CSD prediction method by extracting two-dimensional crystal size information from online microscopic images.