Research

Physics-Grounded AI Surrogate for Bridge Scour and Digital-Twin Decisions

• CFD and flume experiments build the DBscour database
• Bayesian DeepONet predicts full-field scour in seconds
• Reliability analysis links scour risk to bridge safety grades A–E
• Digital-twin alerts with time-to-critical for pre-emptive action

Funded by MOLIT Young Researcher Program (KAIA)

Deep Learning-Based Rapid Storm Surge Prediction

• Deep-learning surrogate of storm-surge simulations
• High-speed, high-accuracy forecasting for coastal cities
• Supports disaster response decisions

AI-Based Urban Cooling Technology Development

• Urban thermal environment analysis
• AI-optimised cooling strategies
• Extreme-heat mitigation planning

Risk-Informed Fire-Resistant Structures and Standards

• Performance-based fire-safety criteria
• Lithium-battery storage warehouses
• Risk-based compartmentation design