Estimating mixture properties from batch distillation using semi-rigorous and rigorous models

Batch distillation experiments on laboratory scale are conducted in order to estimate NRTL parameters describing thermodynamic mixture properties. For this purpose, distillate composition and column temperatures are measured as functions of the batch hold-up and then used to fit model parameters with a non-linear regression approach. Two different models are presented, a semi-rigorous model relying on the assumption of constant molar overflow and a rigorous model from the commercial flowsheet simulator Aspen Custom Modeler. For both models, the NRTL parameters are estimated with good agreement to the experimental data. However, the semi-rigorous model is much simpler and therefore requires far less computational effort.