Projects
CURE Lab — Research Portfolio
We lead interdisciplinary national R&D programs bridging nano-scale rock physics to field-scale deployment — advancing geological carbon storage, AI-driven subsurface modeling, unconventional energy systems, and digital geoscience platforms.
Low-Carbon High-Yield Unconventional Oil Recovery
Develops next-generation thermal recovery technologies for oil sands with substantially reduced carbon footprint. Focuses on hybrid steam-solvent injection processes, intelligent operation control, and geomechanically coupled reservoir simulation to maximize oil production while minimizing steam-to-oil ratio and greenhouse gas emissions.
- Hybrid ES-SAGD and solvent-steam co-injection process simulation
- Thermo-geomechanical coupled reservoir simulation (SAGD + geomechanics)
- Machine learning models for SAGD performance prediction from well log data
- Non-condensable gas (NCG) injection timing and rate optimization
- Well pair spacing, toe-up preheating, and operating strategy optimization
- Economic viability analysis of eco-friendly oil sands development
Cross-Border CCS Network Optimization
Designs an integrated CO₂ capture–transport–storage network linking Korea's major industrial emission sources to offshore and overseas geological storage sites. Develops multi-objective optimization frameworks for hub-and-spoke CO₂ logistics, incorporating economic, safety, and regulatory constraints for a commercially viable CCS supply chain.
- CO₂ source mapping and industrial cluster analysis for hub site selection
- Economic comparison of ship-based vs. pipeline CO₂ transport scenarios
- Integration with blue hydrogen supply chains (Korea–Canada, Korea–Australia)
- Network topology optimization using mixed-integer programming
- Risk, safety, and environmental impact assessment of CO₂ infrastructure
- Policy and regulatory scenario modeling for cross-border CCS frameworks
Nano-Scale Offshore CCS Integrity
Investigates CO₂ containment security and caprock seal integrity for offshore geological storage in Korea's East Sea, spanning nano-scale pore structure through field-scale deformation response. Combines generative AI for digital rock augmentation, coupled thermo-hydro-mechanical (THM) simulation, and satellite InSAR geodetic monitoring to deliver a multi-scale integrity assessment framework.
- SEM/micro-CT pore imaging, segmentation, and pore network extraction of caprock samples
- SinGAN/GAN-based digital rock data augmentation for uncertainty assessment
- THM coupled geomechanical simulation of CO₂ injection and caprock response
- InSAR surface deformation inversion to characterize CO₂ storage reservoirs
- ML-assisted assessment of CO₂ leakage risk through legacy wells and fractures
- Geochemical modeling of CO₂–brine–mineral interactions and trapping mechanisms
Digital Rock Physics + AI Platform
Builds an integrated digital rock physics (DRP) platform that couples pore-scale multi-physics simulation with AI-driven property estimation and uncertainty quantification. Targets automated characterization of reservoir and caprock petrophysical properties — permeability, relative permeability, capillary pressure, and elastic moduli — from 3D CT/SEM images, with full uncertainty propagation from pore to field scale.
- 3D CT/SEM image segmentation and pore network extraction for rock microstructure analysis
- Lattice Boltzmann method (LBM) for pore-scale single- and multi-phase flow simulation
- Deep learning (CNN, PoreFlow-Net) for rapid permeability and porosity prediction
- GAN-based synthetic rock sample generation for data augmentation
- Multi-scale upscaling from pore-scale physics to core- and reservoir-scale properties
- Uncertainty propagation framework for petrophysical property estimation
Generative AI for Geological Media Modeling
Develops generative AI workflows for realistic geological model synthesis, augmentation, and data conditioning. Focuses on single-image and process-based generative approaches that reproduce complex geological heterogeneity — fluvial channels, turbidite lobes, deltaic systems — while honoring sparse well log and seismic observations for robust subsurface uncertainty assessment.
- SinGAN-based geological model augmentation from a single training image
- Conditioning generative models to sparse well log and dynamic production data
- Minimum acceptance criteria for quality evaluation of generative geological models
- Integration of SinGAN with ES-MDA for history matching of CO₂ storage reservoirs
- Probabilistic geological scenario generation for subsurface uncertainty quantification
Generative AI Reservoir Modeling for Gas Storage & CCS
Develops foundation model-based frameworks for 3D reservoir model generation and dynamic data assimilation, targeting underground natural gas storage and CO₂ geological storage applications. Integrates multi-source subsurface data (well logs, 3D seismic, production history, InSAR) into geologically consistent ensemble reservoir models for robust production forecasting and uncertainty quantification.
- Foundation AI model for multi-conditional 3D subsurface model generation
- Ensemble Smoother with Multiple Data Assimilation (ES-MDA) for history matching
- SinGAN/GAN-based geological prior model generation and conditioning
- Integration of well, 3D seismic, BHP, and InSAR deformation data
- CO₂ plume migration prediction and storage performance uncertainty quantification
- Transfer learning for adapting pre-trained models to new reservoir settings
CO₂ Storage Efficiency Enhancement
Developed methods to enhance CO₂ injectivity and storage efficiency in saline aquifers and tight formations. Investigated EGR (Enhanced Gas Recovery) coupled with CCS in the Duvernay Shale, Canada, and optimized multi-well injection scheduling for the Pohang Basin, South Korea.
Commercial Carbon Storage Exploration
Conducted 3D seismic acquisition, interpretation, and geostatistical modeling to evaluate West Sea (Yellow Sea) offshore storage sites for commercial-scale CO₂ geological storage. Developed probabilistic volumetric storage capacity estimates and site ranking frameworks.