State estimation in batch crystallization using reduced population models

This contribution deals with the design of an observer for state estimation of a batch crystallizer, which is described by a detailed population balance model. Therefore, the rigorous model containing (partial) integro-differential equations is first reduced by applying an integral approximation technique to a model of moments that consists of only a few ordinary differential equations. This reduced model then serves as the basis for the design of a Luenberger type observer. The performance of the observer is finally demonstrated in various disturbance scenarios, using the rigorous population balance model for the simulation of the crystallizer plant.