A new correlation model of entrainer properties and process economics for ternary azeotrope separation by extractive distillation

A correlation model was explored and established for the first time, linking the properties of entrainers to the process economics of extractive distillation separation for ternary azeotropes. Firstly, the extraction process of ternary azeotrope (methyl tert-butyl ether/ethanol (EtOH)/water) with common entrainers was optimized using a multi-objective optimization method based on a genetic algorithm. The optimal process parameters were obtained, and dimethyl sulfolone emerged as the best entrainer for the system. An economic correlation model was developed using machine learning to link the properties of entrainers and total annual cost (TAC). The model exhibited a determination coefficient of 0.993 and a single percentage error of less than 2% for each data set, indicating a significant fitting degree and prediction accuracy. The feasibility of selecting the optimal entrainer and calculating the economic benefit of the model was verified through another ternary azeotrope system (i.e., tetrahydrofuran/EtOH/water). The model provides valuable insights into energy savings and entrainer screening in extractive distillation. Additionally, the influence of the heat integration process with different entrainers on the economic benefit was analyzed, and the optimum entrainer changed to glycerol after the heat integration process. This finding suggests that determining the optimum entrainer should consider the possible heat inte-gration process design. Finally, the relationship between the intermolecular interaction mechanism and sepa-ration effect was revealed through quantum chemical calculations.