Integrating rigorous techno‐economic analysis into generative process synthesis for economic optimization
Year: 2026
Type: article
Abstract: This study presents a fully automated dual‐loop framework integrating detailed techno‐economic analysis (TEA) into generative process synthesis. A trained generative model proposes process topologies, while an inner loop combines Aspen Plus equation‐oriented initialization with trust‐region Bayesian optimization (TuRBO) of an external simulation–TEA evaluation. This evaluation includes discrete equipment‐sizing and utility‐selection decisions. Case studies on cyclohexane and ethylene glycol production show how particle swarm optimization (PSO) uses the resulting economic feedback to improve on the best randomly sampled designs. Relative to the best randomly sampled designs, annualized profit increases by 12.2% for cyclohexane, while the annualized deficit decreases by 30.6% for ethylene glycol. These results demonstrate the integration of topology generation, rigorous simulation, and economic evaluation in an automated design workflow.
Source: AIChE Journal
DOI: 10.1002/aic.70758
Institution: Seoul National University
Language: English
Subfield: Control and Systems Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: bronze
Awards: G2473509793, G698531324
OpenAlex ID: W7226401605
