Task-based design of solution crystallization processes: Modelling of tasks for optimization driven process synthesis

Involved people: Prof.ir. J. Grievink

MSc student: Louise Clement

Crystallisation is one of the most important separation processes in industry. The development of industrial crystallisation processes is however hampered by the existing design approach. As a result major fundamental problems like limited flexibility, non-linear behaviour and off spec and varying product quality cannot be addressed and solved adequately. A paradigm shift from equipment-based into phenomena based design is required to address these fundamental problems in industrial crystallisation.

Recently the TU Delft started with the development of a task-based design strategy, which is an example of a phenomena based approach. The crystallization tasks, representing fundamental changes needed to arrive at desired product quality, are used as building blocks rather than existing equipment as is done in current design procedures.

As the task based crystallizer can be built from an arbitrary number of crystallization tasks the design of the integrated crystallization process will be complicated. Therefore a modeling and optimization tool is needed which can be used to analyse quickly the different alternatives. The model system should contain a model library for the different crystallization tasks and should contain a framework to build a network of these crystallization tasks to represent the final crystallization process. In order achieve robustness and allow for the optimization of the process conditions, the model structure and the dependency of the state variables with respect to the process conditions should be analysed upon implementation. The model framework should be implemented in gPROMS.

This MSc thesis project is intended to be a pre-cursor to a more general and comprehensive project on developing a computational multi-scale model for task based crystallisers. It is meant to get a feel for the complexity of the modelling task at hand. The objective is to explore the feasibility of a computational approach to the systematic synthesis of task-based crystallization units in solution processes.