AI-assisted reliability-constrained design and operation optimization of biogas-to-methanol systems under feedstock uncertainty

  • Chung, Hegwon; 
  • Ba-Alawi, Abdulrahman H.; 
  • Kim, Pyung Soon; 
  • Shin, Junghun; 
  • Kim, Jiyong
Citations

WEB OF SCIENCE

0
Citations

SCOPUS

0

초록

Methanol production from biogas has gained attention as a sustainable pathway for utilizing renewable carbon resources. However, its deployment is challenged by inherent feedstock variability, which can adversely affect process stability and efficiency. This study proposes an AI-based reliability-constrained framework for the design and operational optimization of a biogas-to-methanol process. A process model of the integrated reforming and methanol synthesis system was developed to generate performance data. A residual neural network surrogate was then constructed to capture process behavior, achieving high predictive accuracy (R2 ' 0.94). Multi-objective optimization using NSGA-II was applied to identify optimal operating strategies by simultaneously considering unit production cost (UPC), net CO2-equivalent emissions (NCE), exergy efficiency (EXF), and methanol production. The analysis was performed across three process configurations: a baseline configuration (P1), a configuration with external syngas ratio adjustment through material supplementation (P2), and a configuration with internal syngas ratio adjustment (P3). Then, uncertainty analysis was performed using Monte Carlo simulation, and a coefficient of reliability was developed to define reliability-indexed operating bounds for operational optimization. The results show that optimized designs exhibit methanol production rates of 278–346 kg/h, UPC of 0.63–0.76 USD/kg, NCE of − 0.86 to − 0.74 kgCO2/kg, and EXF of 46–50 %. Under 30 days of real plant operation at a target reliability of 98 %, P2 achieves superior robustness, maintaining stable production and reducing cumulative CO2 emissions by 8,300 kg, while P1 exhibits repeated reliability violations. Overall, the framework enables reliable and adaptive operation under feedstock uncertainty, supporting robust biogas-to-methanol production. © 2026 Elsevier Ltd.

키워드

Biogas-to-methanol; Feedstock uncertainty; Operational optimization; Process design; Reliability
제목
AI-assisted reliability-constrained design and operation optimization of biogas-to-methanol systems under feedstock uncertainty
저자
Chung, Hegwon; Ba-Alawi, Abdulrahman H.; Kim, Pyung Soon; Shin, Junghun; Kim, Jiyong
DOI
10.1016/j.enconman.2026.121934
발행일
2026-11-01
유형
Article
저널명
Energy Conversion and Management
권
367