Uncertainty quantification for chromatography model parameters by Bayesian inference using sequential Monte Carlo method

Abstract

Many model-based optimization methods have been proposed for chromatographic processes to ensure product quality and efficiency, but uncertainty of model parameters should be considered to assure robust design and operation. In this study, we developed a sequential Monte Carlo (SMC) parameter estimation method for chromatographic processes to estimate the parameter uncertainty rigorously within a reasonable amount of computation time. As an example, separation of glucose and fructose is considered. Through the example using artificial data, we confirmed that SMC can perform estimations more efficiently than the existing method, Markov chain Monte Carlo. Furthermore, through the example using lab-scale experimental data, we confirm that the time and effort for the sample analysis to identify the concentration of each component can be eliminated. We also examined the relationship between the number of cores and computation time for parallel implementation of SMC.