This paper presents a general approach for the design/selection of single molecularstructures with prescribed functionality, based on mathematical programmingtechniques. The strategy relies on an efficient decomposition of the design problem intotwo levels, relative to the identification of structurally feasible candidate molecules thatsatisfy the property constraints, and their ranking regarding a set of performanceobjectives that include both operational and safety aspects. Group contribution methodsand correlations are employed to model a significant fraction of the physical propertiesused in the problem formulation. The approach introduced is demonstrated byapplication to an example commonly found in aromatics industry – the design ofextraction agents to promote L/L separations.
展开▼
