(371l) Automatic Design and Optimization of Extractive Distillation Sequence for Multicomponent Azeotropic Systems

In chemical production, 40%-60% of energy consumption and equipment costs are related to the separation process. It is estimated that more than 95% of liquid separation in more than 90% of chemical plants and refineries is performed by distillation^[1], so the distillation sequence design is of great significance to reduce energy consumption. There are often azeotropes in the distillation process, which increases the difficulty of the design of the distillation sequence. Such a design is traditionally achieved by the trial-and-error approach based on graphical analysis of experts, which may not be optimal; therefore, a systematic approach to generate the optimal azeotropic distillation sequence is of great importance.

To address this issue, this presentation proposes a novel method for the automatic design and optimization of the Extractive Distillation Sequences (EDS) for the multicomponent azeotropic system. The overall procedure has two phases: (1) the first phase is based on an inexplicit binary-tree enumeration to select the most suitable entertainer and generate several most promising separation sequence candidates^[2]; (2) in the second phase, we incorporate a superstructure which may or may not include preconcentration columns to optimize the EDS candidates obtained previously.

In the first phase, primary structures of candidate separation sequences are first generated based on the adjacency matrix method and the binary tree search algorithm, without considering a specific entrainer; in the binary tree search algorithm, a primary economic evaluation is conducted, and the separation sequences with large costs are filtered out from the candidate pool. Then, the entire separation sequences are completed by considering different entrainers based on the primary structures. Unseparated azeotropes are separated by the suitable entrainers, which are screened through an automated procedure considering the relative volatility and the component compositions. These generated distillation sequences are initial sorted by short-cut calculation of TAC, in which columns are evaluated based on whether it is an extractive distillation column or a regular one. Afterwards, the most promising EDS are passed on to the second phase, in which further optimization are conducted using genetic algorithms and a novel superstructure that considers the possibility of including preconcentration columns. The top-ranked EDSs are rigorously simulated in ASPEN Plus, which validates the proposed short-cut economic evaluation method and the overall design/optimization scheme. The proposed approach is illustrated using two case studies: (1) a four-component system of methyl acetate, methanol, ethanol, and water with four binary azeotropes, and (2) a six- component system of methanol, ethanol, isopropyl alcohol, 2-butanol, water and 1-butanol containing four binary azeotropes.

The proposed approach provides a systematic method for the automatic generation, evaluation and optimization of EDSs considering multiple entrainers. Compared to the traditional approach, this approach is especially useful to seek promising EDSs, when there are multiple azeotropes in the system and the problem becomes highly combinatorial. Case studies show that the proposed method can provide more economical separation sequence compared to some recent studies in literature. It is also found that the optimal EDS designed by empirical rules is no longer applicable for azeotropic systems.

Keywords: Multicomponent azeotropic system; Distillation sequence synthesis; Extractive distillation; Economic evaluation; Superstructure-based optimization

Reference:

[1] Jose Adrian Chavez Velasco, Mohit Tawarmalani, Rakesh Agrawal. Systematic Analysis Reveals Thermal Separations Are Not Necessarily Most Energy Intensive. Joule, 2021, 5(2), 330-343.

[2] Jiaqi Meng, Chao Wang, Yunhai Bai, Yachao Dong, Jian Du. Automatic Design of Extractive Distillation Sequence for a Multicomponent Azeotropic System. Industrial & Engineering Chemistry Research, 2024, 11, 1080.