Answers are grounded in this paper and cite their sources.
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.
No takes yet. Share an insight, caveat, or question.
Son et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: