Process robustness evaluation for various operating configurations of multi-column chromatography processes with nonlinear isotherm

Abstract

This study proposes a method for evaluating process robustness of various operating configurations for multi-column chromatography processes. The robustness of industrial chromatography processes, which refers to the probability of achieving target purity under various uncertainties, is a critical metric. However, quantification of robustness for multi-column chromatography has not been established yet. In this study, we estimate uncertainty in purity as posterior predictive distribution within a framework of Bayesian statistics and quantify the robustness as joint probability.Our proposed method was demonstrated through a comparison of the robustness for five different operating configurations of simulated moving bed (SMB) chromatography processes. The joint probability of product purity revealed that one of the F-shaped configurations, which have been reported in a past study is the most robust, while the conventional SMB configuration is the least. Furthermore, uncertainty of internal concentration profiles was analyzed to investigate influences of internal flow rates on the robustness.