Simultaneous parameter estimation and optimisation of a seeded anti-solvent crystallisation

A population balance incorporating nucleation, growth and agglomeration, solved using quadrature method of moments was coupled with a parameter estimation procedure. The seeded anti-solvent crystallisation of Paracetamol from methanol and water was chosen as the model system. All parameters concerned were regressed from moments calculated using the measured square weighted chord length distribution (CLD) generated by the FBRM. The FBRM and the concentration data are utilised together to obtain experimental moments that reflect the mass of solids in the tank. Using the estimated kinetic parameters, the crystallization model was validated using an additional experiment with a new non linear addition rate. Experimental crystal size distributions measured by laser diffraction are compared to CSDs calculated by the model and found to be in good agreement. No such work exists in the literature using FBRM to model an anti-solvent system which considers agglomeration. Based on the kinetic parameters estimated using the above method, the solution to the optimal anti-solvent addition rate profiles was obtained by applying nonlinear constrained multi-objective free final time formulation optimization on the validated model. These profiles were experimentally tested and CSD were compared with experiments used in the parameter estimation procedure. A 73.3% reduction in batch time was achieved with little impact on the CSD. Analyses of the various conflictions are presented with the aid of a pareto optimal plot to provide the practitioner with increased flexibility.