01
Physics-Informed Graph Learning
Graph neural networks that encode power-grid topology and physical structure for large-scale AC-OPF and topology-changing systems.
Energy AI · Power Systems · Graph Learning
Simple but effective AI
for large-scale power systems.
I develop physics-informed and graph-based AI methods for power-system optimization, foundation models, and renewable-energy forecasting — connecting machine learning with the structure and constraints of real energy systems.
CONNECTING THE DOTS
KS
PhysicsGraphsOptimization
AC-OPF
Grid FM
GNN
Power Systemsdomain-grounded AI
Graph Learningtopology-aware models
Foundation Modelstransfer & generalization
Renewablesforecasting under uncertainty
01
About
AI that respects the grid.
I am a Ph.D. candidate in Electronic Engineering at Sogang University and a member of NICELAB. My research lies at the intersection of energy AI, power systems, graph neural networks, and optimization.
I received my B.S. degree in Electronic Engineering from Tech University of Korea in 2021. In 2025, I worked as a visiting graduate student in the Mathematics and Computer Science division at Argonne National Laboratory, where I contributed to physics-informed graph learning and grid foundation-model research for AC optimal power flow.
02
Research
Three threads, one research direction.
Building scalable, transferable, and physically meaningful AI for modern power systems.
02
Grid Foundation Models
Reusable grid representations for multi-topology pretraining, transfer learning, zero-shot adaptation, and feasibility-aware scientific AI.
03
Renewable Energy Intelligence
Spatiotemporal forecasting and robust learning for renewable generation, including large-scale PV systems with severe missing data.
03
Selected Publications
Recent work.
A selection spanning power-flow acceleration, grid foundation models, graph learning, and energy forecasting.
2026
LUMINA: A Grid Foundation Model for Benchmarking AC Optimal Power Flow Surrogate Learning
Hongwei Jin, Keunju Song, Zeeshan Memon, Yijiang Li, Stefano Fenu, Hongseok Kim, Liang Zhao, Kibaek Kim
2026
Physics-Informed Graph Learning Acceleration for Large-Scale AC-OPF with Topology Changes
Keunju Song, Kyungnam Park, Sua Choi, Seunguk Kim, Tae-un Kim, Youngmin Choi, Sang-Won Min, Hongseok Kim
2026
SABLE: GPU-Based Power Flow Accelerator for Sparsity-Aware Batched Learning
Suho Park, Keunju Song, Hongseok Kim
2026
LUMINA: Foundation Models for Topology Transferable ACOPF
Yijiang Li, Zeeshan Memon, Hongwei Jin, Stefano Fenu, Keunju Song, et al.
2024
Graph-based Large Scale Probabilistic PV Power Forecasting Insensitive to Space-Time Missing Data
Keunju Song, Minsoo Kim, Hongseok Kim
2022
DTTrans: PV Power Forecasting Using Delaunay Triangulation and TransGRU
Keunju Song, J. Jeong, J. H. Moon, S. C. Kwon, Hongseok Kim
04
Projects
From models to real grids.
Selected research and applied projects across grid optimization, forecasting, and energy management.
2025 —NRF Korea
National Power Grid Optimization with Hardware-Accelerated Physics-Informed Neural Networks
Physics-informed learning and acceleration for large-scale national power-grid optimization.
Sogang University
2024 — 2025KPX
AI for AC Optimal Power Flow
Learning-to-optimize methods for AC-OPF using large-scale, real power-system operating data.
Sogang University
2024Competition
HEFTcom24 — NICE_Forecast
Hybrid energy forecasting and trading competition; 11th overall and 3rd among student teams.
2023LG Electronics
Home Energy Storage Charge / Discharge Control
Scheduling and control research for residential energy-storage systems.
Sogang University
05
Experience
Research across academia and national labs.
2025.09 — 2025.11
Argonne National Laboratory
Visiting Graduate Student · Mathematics and Computer Science Division
- Physics-informed heterogeneous graph learning for AC optimal power flow.
- Hierarchical federated-learning research and collaboration around grid foundation models.
- Contributed to the Argonne GridFM / LUMINA research direction.
2021 — Present
Sogang University · NICELAB
Ph.D. Candidate · Electronic Engineering
Research on energy AI, large-scale power-system optimization, graph neural networks, and renewable-energy forecasting.
2021
Tech University of Korea
B.S. · Electronic Engineering
06
Recognition & Service
Awards, patents, and reviewing.
Awards
2025
Best Graduate Student Award
Sogang Ricci Engineering Academic Award
2024
The Grand Prize
Sogang University Best Paper Award
2023
Excellence Award
Qualcomm & Sogang University Best Paper Award
2022
Encouragement Award
Qualcomm & Sogang University Best Paper Award
Patents
Operating-state determination and learning method for AC power systems
Power-generation forecasting with large-scale missing-data imputation
Split-learning-based PV forecasting for energy privacy
Renewable-energy generation forecasting using multiple weather stations
Reviewer Activities
IEEE Transactions on Sustainable Energy
IEEE Transactions on Smart Grid
IEEE Power Engineering Letters
Nature Communications
Let’s connect
Interested in energy AI,
power systems, or graph learning?
I’m always open to research conversations and collaboration.