A generic and systematic model-based framework for the design of a process monitoring and control system to achieve the desired crystal size distribution (CSD) and crystal shape for a wide range of crystallization processes has been developed. This framework combines a generic multi-dimensional modelling framework, tools for design of set point profiles, for design of PAT (Process Analytical Technology) systems as well as option to perform the uncertainty and sensitivity analysis of the PAT system design. Through this framework, it is possible for a wide range of crystallization processes to generate the necessary problem-system specific model, the necessary set point using the extended analytical CSD estimator and the response surface method (RSM) and a PAT system design including implementation of monitoring tools and control strategies in order to produce a desired product with its corresponding target properties. In addition the impact and influence of input uncertainties on the predicted PAT system performance can be quantified, i.e. the risk of not achieving the target specifications of the crystal product can also be investigated. The application of the systematic model-based framework is divided into three sections: a) the application of the generic multi-dimensional modelling framework are highlighted: i) the capability to develop and further extend a batch cooling crystallization model is illustrated through a paracetamol case study, supplemented by a sucrose crystallization example to demonstrate how the framework supports smooth switching between chemical systems with a minimum modelling effort; ii) a potassium dihydrogen phosphate (KDP) case study is used to demonstrate how the model complexity can be increased, that is, by switching from a one-dimensional to a twodimensional description; b) the systematic framework is used in a case study to design a monitoring and control (PAT) system for a potassium dichromate and KDP crystallization processes to achieve the desired target CSD respectively; and c) Based on the PAT system design in b), the application of uncertainty and sensitivity analysis is then highlighted for the potassium dichromate and KDP crystallization process both in open-loop and closed-loop operation. In the case study, the impact of input uncertainties related to parameters of the nucleation and the crystal growth model on the predicted system performance has been investigated for a one- and two-dimensional CSD and it shown the PAT system design is reliable and robust under considered uncertainties.
Main Research Area:
Gernaey, Krist, Sin, Gürkan, Gani, Rafiqul
Technical University of Denmark, Department of Chemical Engineering, 2012