Abstract

The design of distillation processes must satisfy several requirements – controllability, safety, operability, and economic productivity – that are determined from experimentation, mathematical modeling, and computational simulation. This work investigated the relation between column variables and the mole fraction of methanol (MeOH) in the distillate stream (xD) on a modified Oldershaw column under semi-automatic control to capture column operating behavior. A positive correlation between the MeOH xD and reflux ratio was found. A rigorous RadFrac model from the Aspen Plus V10 software simulated an analogous relationship. The percent mean absolute error (MAE) between the model predictions and experimental observations was 1%. A correlation between the feed flow rate and the MeOH xD was also discovered with a comparison between Aspen Plus predictions and empirical measurements that resulted in a percent MAE of 3%. This information was utilized to design a feasible economic solution for the separation of a 50:50 MeOH isopropanol mixture. This solution consisted of sequential distillation columns. It was found that salaries and utilities were the dominant contributor to total liabilities, a result of transient operating hours. This study demonstrated how merged experimental and model information can be integrated and used to gain access to insights into the design of complex design projects.

Keywords: distillation simulation; distillation experiment; distillation model; distillation design; distillation economics; distillation aspen plus.