Deep Learning-Assisted Flash Calculations for Underground Hydrogen Storage

Jun 1, 2026·
Zhilei Han
,
Bicheng Yan
· 1 min read
Type
Publication
SPE Europe Energy Conference and Exhibition
publications

This work combines conventional thermodynamic flash calculations with deep-learning-assisted phase stability and initialization to improve computational efficiency for compositional underground hydrogen storage simulation.

Authors
PhD Candidate | Reservoir Simulation | Formation Evaluation & Petrophysics

I am a PhD candidate in Energy Resources and Petroleum Engineering at King Abdullah University of Science and Technology (KAUST). My doctoral research focuses on numerical modeling of coupled compositional multiphase flow and microbial reactive transport for underground hydrogen storage (UHS), with emphasis on thermodynamics, reservoir simulation, and computational efficiency.

Before my PhD, I worked for seven years at CNOOC Tianjin as a formation evaluation and logging engineer. This combination of field-scale formation evaluation experience and reservoir simulation research shapes my interest in integrated subsurface characterization, modeling, and decision support.