This paper presents a model for reversed-phase purification of insulin from desamido insulin. The system is described by a reaction dispersive model with a competitive Langmuir isotherm. A model building and calibration method is presented and the model's region of validity is defined. The model is calibrated using only two-component experiments on the raw mixture by the inverse method and then experimentally validated. The model is then used to optimize the system's production rate with both purity and yield requirements. The yield requirement is varied between 80 and 95% to study the effect on the production rate and the operating point. The operating points found with the optimization were found outside the model's region of validity, but the experimental validation of the operating points shows that the model can be extrapolated to the interesting operating points. (c) 2007 Elsevier B.V. All rights reserved.