Deep neural network-assisted fast and precise simulations of electrolyte flows in redox flow batteries

Zixiao Guo, Jing Sun, Shuaibin Wan, Zhenyu Wang, Jiayou Ren, Lyuming Pan, Lei Wei, Xinzhuang Fan and Tianshou Zhao

Applied Energy, 2025, vol. 379, issue C, No S0306261924022931

Abstract: Flow fields are a key component in redox flow batteries, which is to distribute electrolytes onto electrodes at the maximum uniformity with the minimum pump work. Achieving this design goal requires accurate simulations of electrolyte flows and identification of the dead zones where the flows become weak or stagnant. However, conventional case-by-case numerical simulation requires significant computational resources. In this work, we use deep learning to predict the electrolyte flow in flow batteries with a neural network knows as U-Net. The U-Net is well trained by learning the mapping between the input (flow field geometry) and output (velocity magnitude distribution). Results show that the pixel-wise comparison of the velocity magnitudes between the U-Net-predicted results and finite element simulated results exhibits an average Euclidean distance of 442.6 and an average R2 of 0.979, indicating that the electrolyte distribution can be accurately simulated based on the geometric characteristics of flow fields. In addition, dead zones are precisely identified by labeling the regions with low velocity magnitudes. Modifying the channel depth in these regions substantially enhances the under-rib convection, thereby improving the system efficiency by 5.5 % at 200 mA cm−2. Furthermore, compared to the numerical simulation, the U-Net-assisted prediction significantly reduces the computational time by 99.9 %. It is anticipated that the U-Net-assisted simulation provides an accurate and efficient tool for obtaining the velocity distribution of flow fields that can assist the flow field design especially in large quantities and large scale.

Keywords: Redox flow battery; Flow field design; Deep neural network; Electrolyte distribution; Computational cost; Dead zone identification (search for similar items in EconPapers)
Date: 2025
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0306261924022931
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:appene:v:379:y:2025:i:c:s0306261924022931

Ordering information: This journal article can be ordered from
http://www.elsevier.com/wps/find/journaldescription.cws_home/405891/bibliographic
http://www.elsevier. ... 405891/bibliographic

DOI: 10.1016/j.apenergy.2024.124910

Access Statistics for this article

Applied Energy is currently edited by J. Yan

More articles in Applied Energy from Elsevier
Bibliographic data for series maintained by Catherine Liu ().