The DGYM research group at King Abdullah University of Science and Technology (KAUST) held its Annual Conference on September 28, 2025. The one-day symposium brought together researchers and industry professionals to share recent progress in numerical simulation and AI-enhanced modeling for subsurface energy applications.

Connecting Academia and Industry

The conference adopted a hybrid format, welcoming both on-site and online participants. Representatives from Sinopec attended in person at KAUST, while participants from Saudi Aramco, Zhenhua Oil, and other organizations joined virtually. This format facilitated exchange between academic researchers and industry practitioners working on carbon storage, hydrogen storage, geothermal energy, and oil and gas recovery.

Prof. Bicheng Yan opened the conference with a welcome speech, introducing the day's program and the symposium's goals for fostering collaboration.

Technical Program

The symposium was organized into three sessions throughout the day:

Session 1: Molecular & Pore-scale Simulation

The morning session covered fundamental studies at molecular and pore scales:

• Dr. Arun Narayanan Nair discussed molecular dynamics simulations investigating salt effects on CO₂ geological storage
• Xinyu Yao presented work on cushion gas effects in underground hydrogen storage
• Dr. Remya Nair shared research on nanographenes as anode materials for metal-ion batteries
• Haotian Li described deep learning approaches for pore structure characterization and permeability prediction in shales

Session 2: Reservoir Simulation Development

The late morning session focused on reservoir simulation tools and methods:

• Zhilei Han presented a numerical scheme coupling compositional flow with microbial reactive transport for hydrogen storage
• Huseyn Yusifov introduced GYMS-Geothermal, a Julia-based reservoir simulator framework
• Billal Aslam discussed the Coarse-Grid Network Model for accelerated reservoir simulation
• Biao Zhou described an improved embedded discrete fracture model for well-test analysis

Session 3: Deep Learning for Reservoir Simulation

The afternoon session explored machine learning applications:

• Dr. Junjun Chen presented deep learning methods for inverse modeling and monitoring design
• Haotian Li discussed multiscale reservoir simulation using super-resolution techniques
• Ziyou Liu shared work on optimization of coupled geothermal systems
• Qirun Fu introduced a hybrid CNN-Transformer model for CO₂ plume monitoring
• Zhen Li described a machine learning framework for wellbore trajectory prediction and control

The day concluded with a panel discussion, providing an opportunity for questions and informal exchange among participants.