(185v) A New Optimization-Based Computer-Aided Molecular and Mixture Design (OptCAMD) Framework

Chemical engineering has been expanding its focus from process-centered products such as ethanol and ammonia to chemical-based products such as mosquito repellent sprays and blended fuels [1]. Modern society requires many chemical-based products for its survival, such as fuels, plastics, detergents, medicines and foods [2]. The sustainability of modern society depends on the continuous availability of these products. Nowadays, chemical-based product design and development has become a key topic in chemical engineering since high value-added products, green chemistry and product-process sustainability have started to receive increasing focus in research and development. Various requirements have to be satisfied for different product design problems, which are hard to be determined from limited number of experiments, considering the cost of time and money. At the same time, potential products may be missed with empirical and knowledge-based design methods. Therefore, it is important to develop a systematic way to determine the optimal products based on the appropriate product attributes that need to be converted to property constraints. The design of molecular products is the foundation of all chemical-based product design problems. Even with established molecular design methods, the mathematical formulation and solution of the design problem for a wide range of chemical-based products is still a tough problem, considering the diversity of products and their requirements, huge number of isomers, and the availability of models for the prediction of the needed properties. Therefore, it is necessary to developed new molecular and mixture product design tools.

In this paper, an optimization-based framework for molecular and mixture product design is presented. The molecular product design problem is decomposed into preliminary design, isomer generation and evaluation and verification steps in this framework. In the preliminary design step, an optimization-based mathematical programming model is established to generate feasible sets of molecular groups. In the isomer generation step, a SMILES-based isomer generation strategy is proposed to generate all isomers. In the evaluation and verification step, the properties of the isomers are verified using database and software with rigorous property model, to evaluate the availability of the generated molecules and accuracy of the properties, and further requirements including unconventional properties and process constraints are then evaluated. In the case of mixture design, a MILP/MINLP problem is formulated to find the optimal mixture (chemical identity and their composition) that satisfy a set of desired target properties. Typical examples, are tailor-made fuel blends. The design work-flows for molecular design and mixture design have been implemented into a chemical-based product design tool “OptCAMD”, which is available as an option in ProCAPD [3], the chemical product design tool. Several design case studies are presented to illustrate the application of OptCAMD involving molecular and mixture design problems.

References

[1] Zhang, L., Fung, K. Y., Zhang, X., Fung, H. K., & Ng, K. M. (2017). An integrated framework for designing formulated products. Computers & Chemical Engineering, 107, 61-76.

[2] Zhang, L., Babi, D. K., & Gani, R. (2016). New vistas in chemical product and process design. Annual Review of Chemical and Biomolecular Engineering, 7, 557-582.

[3} Kalalul, S., Zhang, L., Fang, Z., Choudhury, H. A., Intikhab, S., Elbashir, N., Eden, M. R., Gani, R. (2018). Computer-aided chemical product design – ProCAPD & tailor-made blended products. Computers & Chemical Engineering. DOI: 10.1016/j.compchemeng.2018.03.029.