Deterministic Global Optimization of Multistage Melt Crystallization Processes in Hydroformylation

This paper deals with the deterministic global optimization of multistage melt crystallization processes with application to mixtures of n-/iso-aldehydes resulting from hydroformylation. For this purpose we adopt a process model including experimentally determined parameters from the literature. We extend that model from commonly used countercurrent cascades to arbitrary process configurations and reformulate it for better computational performance. The resulting model is a MINLP due to operational and structural degrees of freedom. Global optima for the process costs with respect to an economical cost function are determined using BARON/GAMS. The results show that there are new process configurations outperforming the conventional countercurrent cascade, depending on parameter values.