(568b) The inside Story of an Extractive Distillation Column: Process Modelling Based Scale-up with Plant Data in a Debottlenecking Case Study

Distillation is the primary separation process used in petrochemical and refining industry. Sometimes it is technically difficult to separate close boiling compounds using normal distillation and it is economically expensive as well. Some of these close boiling compounds can be separated using extractive or reactive distillation columns. Extractive distillation, reactive distillation plays an important role by combining two unit operations in a single equipment to reduce CAPEX and OPEX.

In extractive distillation, a solvent is added to alter the boiling points of the compounds to be separated. Modelling of these advanced distillation processes is very important to optimize the operation as well as to retrofit and/or debottleneck the existing plants. The interaction between solvent and the components in the original mixture plays an important role in the extractive distillation. Thermodynamic vapour liquid equilibrium (VLE) models are used to describe these interactions. These VLE models use binary interaction parametes (BIP’s) to quantify these interactions that will affect the boiling points of the compounds.

This paper explores the methodology of process modeling based scale-up by regressing the binary interaction parameters from laboratory as well as plant data. By using these parameters, the accuracy of the plant model was improved further and it is used for the debottlenecking study of a petrochemical plant.