An integrated multiphysics modeling and ANN–GA optimization framework for high-performance proton exchange membrane water electrolyzer stacks
14
item.page.dc.date.issued
2026-05-01
item.page.dc.type
item.page.dc.identifier.eissn
25901745
item.page.dc.identifier.doi
item.page.dc.identifier.scopus
2-s2.0-105034207769
item.page.oaire.citation.title
Energy Conversion and Management X
item.page.oaire.citation.volume
30
item.page.dc.rights.holder
SCOPUS
item.page.dc.identifier.citation
Energy Conversion and Management X Vol.30 (2026)
punsarn.item.page.citation
Suvanjumrat C., Rumnum S., Seanglumlert C., Rattanaphan S., Priyadumkol J., Promtong M., Punurai W., Boekfah A. An integrated multiphysics modeling and ANN–GA optimization framework for high-performance proton exchange membrane water electrolyzer stacks. Energy Conversion and Management X Vol.30 (2026). doi:10.1016/j.ecmx.2026.101801 Retrieved from: https://repository.li.mahidol.ac.th/handle/123456789/115968
item.page.dc.title
An integrated multiphysics modeling and ANN–GA optimization framework for high-performance proton exchange membrane water electrolyzer stacks
item.page.dc.contributor.author
item.page.oairecerif.author.affiliation
item.page.dc.contributor.correspondence
item.page.dc.contributor.other
item.page.dc.description.abstract
Enhancing the performance of proton exchange membrane water electrolyzer (PEMWE) stacks necessitates an integrated modeling and optimization framework capable of simultaneously delivering high predictive fidelity, optimization robustness, and design flexibility—capabilities that are typically addressed in isolation in existing studies, thereby constraining system-level applicability. In this work, a comprehensive multiphysics PEMWE stack model is developed and implemented in MATLAB 2024a, enabling accurate representation of coupled electrochemical, thermal, and operational phenomena. The model demonstrates strong predictive performance, achieving an average absolute deviation below 4.36% when validated against experimentally measured polarization curves across a broad range of operating conditions and parameter combinations. To facilitate efficient and reliable optimization, a unified artificial neural network–genetic algorithm (ANN–GA) framework is established, with water flow rate (Q<inf>H<inf>2</inf>O</inf>), operating water temperature (T<inf>H<inf>2</inf>O</inf>), active area (A), number of cell stacks (n<inf>Stack</inf>), and current density (J) selected as key decision variables to enhance cell voltage and power density. A comprehensive training dataset is generated using the validated multiphysics model to ensure thorough coverage of the design space, while sensitivity analysis identifies the number of cell stacks as the dominant parameter governing overall PEMWE stack performance. The ANN–GA model exhibits exceptional predictive accuracy, yielding a coefficient of determination (R<sup>2</sup>) of 0.99905 and a mean squared error of 315.48. Evaluation of the optimized parameter set using the physics-based PEMWE model results in a mean prediction error below 10.23%, confirming strong consistency between data-driven optimization outcomes and first-principles performance predictions. Overall, the proposed integrated multiphysics modeling and intelligent optimization framework provides a robust, interpretable, and computationally efficient platform for steady-state PEMWE stack analysis and optimization, offering valuable guidance for advanced system design, operational strategy development, and next-generation hydrogen production technologies.