Dynamics and MPC of an Evaporative Continuous Crystallization Process

Abstract

Abstract Only few applications of Model Predictive Control (MPC) for continuous crystallization processes are found in the open literature. A reason for this may be the nonlinear behaviour of this process. This work applies the MPC approach to improve the control of a Mixed Suspension – Mixed Product Removal (MSMPR) evaporative crystallizer. The modelling of the process was developed for potassium sulphate crystallization due to its well established kinetics for nucleation and growth. The model is described by six equations: the Population Balance Equations (PBE), integrated in moment equations; solute mass balance and the energy balance. The main objective was the regulatory control for the most common disturbances in the process: the feed solute concentration and the feed flowrate. The considered manipulated variables were temperature and steam outflow rate. The results showed that MPC is a feasible control strategy for the process. This is observed especially when the main interest is keeping under regulatory control more easily measured output variables, such as the second order momentum (related to the superficial area of crystals), the concentration, and the suspension temperature. The MPC controller was also compared to PI controllers, presenting better performance and desirable zero offsets for the outputs of greater interest.