Sensitivity analysis of pareto solution sets of multiobjective optimization for a batch cooling crystallization process

An increased interest in pharmaceutical crystallization has been observed in recent years where the product quality is affected by variations in product crystal size distribution, purity, and morphology of the crystals which in turn affect the downstream processes such as filtration, milling, and eventually, formulation. Therefore, it is necessary to ascertain the optimal operation of the process in terms of optimal temperature policies in order to achieve the desired product crystal size distribution. In this work, the optimization of an unseeded batch cooling crystallization of paracetamol within a multi-objective framework is studied. The weight mean size and coefficient of variation of final crystals are considered as objectives, which give rise to a set of optimal solutions known as Pareto-optimal solutions using non dominated sorting genetic algorithm. The present work focuses on evaluating the sensitivity of optimal solutions for induced disturbances and uncertainties in process model mismatch during practical applications through simulation and optimization. It has been observed that the selection of a cooling profile exerts a strong influence on final product crystal size distribution which reiterates the importance of optimization and control in batch cooling crystallization processes industrially at a large scale. Further, in this study, an initial attempt has been made to provide some directions regarding solution approaches to the sensitivity issues.