Inherent flexibility design strategy of extractive distillation for binary azeotropic separation

Abstract

The steady-state design and control studies of extractive distillation (ED) for azeotropic separation has received significant interest in the recent years. These studies however are usually conducted sequentially where the steady-state design is first developed followed by conducting the control analysis. In contrast to existing studies, this work incorporated the operational flexibility into the steady-state design of the ED so that any external fluctuations can be accounted in the early stage of the design. Two case studies are used to exemplify the inherent flexibility design strategy, i.e., the IPA dehydration using dimethyl sulfoxide (DMSO) and ethylene glycol (EG) as entrainer. Each case study is optimized using two different objective function individually, i.e., the total annual cost (TAC) and the flexibility index (FI). Optimizing both cases using TAC as objective function provides a lower TAC relative to the base case, which aligned with existing studies. In contrast, optimizing both cases using FI enables a greater operational flexibility, but such benefit is always accompanied by the increase in the TAC. During the FI optimization, four different uncertainties are considered, i.e., A, B, C, and D, with a combination of ± 20% feed flowrate and IPA feed concentration. For both cases, the operational flexibility limitation was found to lie in the direction of vertex D when the system is experiencing an increase in feed flowrate and a decrease in IPA feed composition. Overall, it was found that using EG provides 2.3 times better operational flexibility in comparison to that using DMSO, as indicated by the higher FI, while using DMSO provides a lower TAC by about 9%. Here, it was also found that using EG provides a lower trade-off in an increase in TAC over using DMSO, when the FI is incorporated into the steady-state design of the ED.