Computer-aided biocompatible solvent design for an integrated extractive fermentation–separation process

Abstract

Computer-aided process/solvent design is introduced to find a feasible biocompatible solvent for an extractive fermentation and separation process. The designed biocompatible solvent serves as both the extractant for extractive fermentation and the entrainer for extractive distillation, to yield water-free ethanol. Several goals, such as maximizing production rate and extraction efficiency, and limiting solvent utilization, are simultaneously considered in the optimal solvent design problem. Thus, the design was formulated as a mixed-integer nonlinear optimization problem. A two-phase computational scheme was introduced to solve the problem. The mixed-integer hybrid differential evolution (MIHDE) algorithm was first applied to solve the problem in order to obtain a feasible solution. The feasible solution was then served as an initial starting point for the mixed-integer sequential quadratic programming (MISQP) solver to numerically confirm that the optimal design was achieved. We have compared the crisp and fuzzy approaches to the design problem. The fuzzy goal attainment approach is able to address goal trade-offs and yield an overall satisfactory grade for the problem.