Deep learning-based surrogate modeling for underground hydrogen storage

Shuojia Fu

doi:10.1016/j.advwatres.2025.105014

Deep learning-based surrogate modeling for underground hydrogen storage

2025-06-02

0

PRE

AI

DOI：10.1016/j.advwatres.2025.105014

Original

Original request for help

Save

Abstract

En 中文

• We apply Swin-Unet for surrogate modeling of underground hydrogen storage. • Swin-Unet enables accurate and efficient prediction of the spatiotemporal evolution of reservoir pressure and hydrogen saturation. • Swin-Unet improves pressure accuracy and reduces saturation training cost compared to U-Net and Segmentation Transformer.

Deep learning-based surrogate modeling for underground hydrogen storage

Abstract

Journal

Organization