Modeling and optimization of an integrated system for the green dimethyl ether production: Process simulation, modeling with artificial intelligence, and optimization methods

Abstract

Green dimethyl ether is emerging as a promising alternative fuel due to its environmental benefits and potential integration into carbon-neutral energy systems. This study presents a comprehensive workflow for modeling, simulation, and optimization of green dimethyl ether production using an integrated approach that combines artificial neural networks and multi-objective optimization. Process simulation in Aspen Plus was employed to model key units, including methanol synthesis, CO₂ capture, desalination, electrolysis, and cogeneration. Artificial neural networks were trained to predict system performance, achieving high accuracy across multiple process units. A comparative analysis of deterministic, Bayesian, and metaheuristic optimization approaches showed that deterministic optimization offered the best balance between economic feasibility and energy efficiency. The optimized configuration achieved a green dimethyl ether production rate of 3035.7 kg/h, reducing total annual costs and energy consumption. These findings highlight the effectiveness of AI-driven optimization in enhancing sustainable fuel production. The proposed methodology contributes to the transition toward a low-carbon economy by improving green dimethyl ether process efficiency and economic viability.