Pressure-Swing Distillation for Separating Pressure-Insensitive Minimum Boiling Azeotrope Methanol/Toluene via Introducing a Light Entrainer: Design and Control

In this paper, pressure-swing distillation (PSD) is applied to separate pressure-insensitive minimum boiling azeotrope methanol/toluene via introducing chloroform as a light pressure-swing entrainer. The feasibility of this process is verified by analyzing the residue curve maps. Both partially and fully heat-integrated PSD processes are implemented. The steady-state designs of these processes are optimized based on total annual cost (TAC). In this extended PSD process, the high-pressure column (HPC) serves as the homogeneous azeotropic distillation column, while the low-pressure column (LPC) serves as the entrainer recovery column. It is found that the HPC has large temperature slope in the stripper section while with relatively flat trend above. This unique phenomenon makes it beneficial to reduce the energy consumption by preheating the feed and recycling stream using low-grade heat. Both conventional and partially heat-integrated extractive distillation processes with aniline as the entrainer are also implemented to compare the economy with these PSD processes. The results reveal that TAC of the partially heat-integrated extractive distillation process is higher than the conventional one. It also reveals that the partially and fully heat-integrated PSD processes have almost similar TAC and energy consumption, while the optimal PSD process has 5.39% reduction of TAC and 8.32% energy savings compared to the optimal conventional extractive distillation process. The dynamic controllability of this extended PSD is investigated by introducing ±20% disturbances of feed flow rate and composition for both partially and fully heat-integrated processes. The results reveal that the enthalpic state of the feed has crucial influence on the dynamic controllability, as it can influence the temperature profile and composition profiles of HPC greatly. Although the economic optimal PSD process shows weak controllability, robust control can be achieved under another superior design.