An advanced optimization strategy for enhancing the performance of a hybrid pressure-swing distillation process in effective binary-azeotrope separation

The separation of binary azeotropes with low cost and energy consumption is important and challenging for the production of chemicals. A hybrid heat-integrated pressure-swing distillation process (HHIPSD) with heat pump was proposed for the separation of both maximum-and minimum-boiling azeotropes in this work. Compared to the conventional fully heat-integrated pressure-swing distillation process (FHIPSD), the HHIPSD process showed lower total annual cost and higher thermodynamic efficiency in the separation of seven minimum-boiling azeotropes of benzene-ethanol, methanol-acetone, methanol-ethyl acetate (EA), ethanol-EA, acetonitrileethanol, isobutanol (IBA)/isobutyl acetate (IBAc) and diisopropyl ether (DIPE)-isopropyl alcohol (IPA), and two maximum-boiling azeotropes of water-ethylenediamine (EDA) and acetone-chloroform. For the best performance of HHIPSD process, robustness and algorithm optimization were studied in this paper. Scenario analysis method was proven to greatly enhance the convergence of HHIPSD models. The improved genetic algorithm (GA) was modified by particle-swarm optimization and pattern search method and validated by 13 benchmark functions. Meanwhile, different constraint handling methods were studied to ensure the better performance of the improve GA method. With the improvement of model robustness and optimization algorithm, the HHIPSD process achieved 21.6 % and 4.2 % decrease of TAC for water-EDA and methanol-acetone separation, respectively, compared to FHIPSD process. The proposed HHIPSD process together with the optimization strategy suggests the pressure-swing distillation technology can be highly effective in azeotrope separation.

Automated summary

The hybrid heat-integrated pressure-swing distillation process (HHIPSD) with heat pump is proposed for the separation of binary azeotropes, achieving lower cost and higher efficiency. The robustness of the model and algorithm optimization are studied, leading to the improved performance of the HHIPSD process.