(269d) A Computational Methodology for Designing Mixtures

As trial-and-error methods are becoming increasingly intractable with the growing library of chemical compounds, computer-aided molecular design (CAMD) is a necessary tool for the future of chemical production.  CAMD combines optimization techniques with many property prediction models, allowing for a rapid and accurate search through all chemical structures, even those for which no data is available.  These techniques work well for designing single molecules, but a more difficult –and more applicable – problem arises when attempting to design mixtures [1]. 

We propose a novel mixture design approach which couples molecular design with derivative-free and algebraic model-based optimization.  We use derivative-free optimization (DFO) to search in the space of compound properties.  Requirements on properties of the individual components are then converted into molecular structures via our software AMODEO and a recently developed molecular design methodology [2], which relies on Group Contribution methods for property estimation [3].  Finally, we use mixing rules and algebraic global optimization in order to optimize over the composition space for a given set of mixture components.  The approach exploits a key insight to the molecular mixture design problem, namely that the small number of degrees of freedom in the property space can be dealt with efficiently via derivative-free optimization (DFO) algorithms [4].  Several case studies are presented to illustrate the wide applicability of the proposed methodology.

References:

1. Conte, E., R. Gani, and K. M. Ng, Design of formulated products: A systematic methodology, AIChE Journal, 57, 2431-2449, 2011
2. Samudra, A. and N. V. Sahinidis, Optimization-based framework for computer-aided molecular design, AIChE Journal, accepted, 2013
3. Marrero, J. and R. Gani. Group-contribution based estimation of pure component properties, Fluid Phase Equilibria, 183–184:183–208, 2001
4. Rios, L. M. and N. V. Sahinidis, Derivative-free optimization: A review of algorithms and comparison of software implementations, Journal of Global Optimization, DOI 10.1007/s10898-012-9951-y, 2012