Modeling and Calibration of Preparative Chromatography

Sammanfattning: In my thesis, I have modeled adsorption chromatography and simulated, calibrated, and analyzed models of adsorption chromatography. While my main focus has been modeling and calibration of the models by estimating the parameters, the models used have been partly new. The simulations are differentiated from the modeling in that the modeling describes the partial differential algebraic equation, while the simulation describes the discretization and solution of the model. When a model is calibrated, the model parameters are estimated to replicate experimental data. While there are many methods to do this sequentially, the method I have used is the inverse method, where all parameters are estimated simultaneously. It is sometimes harder to converge, but potential errors in the estimation do not propagate. I have developed a method to find how model parameters may depend on experimental errors. The process I have worked most with is ion-exchange chromatography, operated in bind and elute mode, and the model I have used most to describe the adsorption process is the steric mass action model. The calibrated model can then be used to find potential process parameters by optimization. The model can also be used to analyze how the process reacts to different kinds of variations.