Conceptual process and surrogate optimization of acrylonitrile production from glycerol via green propylene

• Published in: Brazilian journal of chemical engineering 2024-09, Vol.41 (3), p.901-910
• Main Authors: Guerra, Gabriel N., Neto, Guilherme J. Musse, Alves, Rita M. B., Pontes, Luiz A. M.
• Format: Article
• Language: English
• Subjects: Chemistry; Chemistry and Materials Science; Industrial Chemistry/Chemical Engineering; Original Paper
• ISSN: 0104-6632; 1678-4383
• DOI: 10.1007/s43153-023-00387-y

Acrylonitrile is a commodity currently produced from petrochemical propylene, with significant economic importance. The development of a sustainable and economically viable process for acrylonitrile production from a renewable resource will contribute significantly to the rise of this technology. This work proposes a conceptual process for producing green propylene from glycerol hydrogenation, followed by producing renewable acrylonitrile from green propylene. The main innovation of the proposed process is the hydrogenation section, where renewable propylene is obtained by glycerol hydrogenation and recovered by absorption and distillation. The proposed process was simulated, and the equipment purchase and operating costs were estimated. The results show that most of the plant costs are related to fractionation operations (45.84% of the purchase costs and 63.84% of the operational costs), among which the final propylene purification column is the most costly. In this way, this column was used as study case for surrogate-based optimization tests, where three surrogate models (polynomial, kriging and artificial neural network) and two optimization methods (quasi-Newton and genetic algorithm) were compared. These tests indicated that an artificial neural network (ANN) and a quasi-newton algorithm were the best combination of surrogate model and optimization algorithm, which resulted in a more significant annual cost reduction for the study case.