Research Topic
Physics-Guided Machine Learning for Operational Risk Assessment under Management-of-Change Conditions in Energy Systems
As industrial chemical and energy systems become increasingly complex and continuously undergo Management-of-Change (MOC) events, developing adaptive and intelligent process safety frameworks has become critically important for preventing fault propagation, protecting assets and personnel, and reducing major economic losses. My research focuses on developing physics-guided machine learning (PGML) frameworks for intelligent process monitoring and dynamic operational risk assessment in chemical processes, oil and gas systems, and refineries. The work integrates machine learning, process safety, and physics-guided model architectures using Signed Directed Graphs (SDGs) to capture changing process topology, connectivity, and causality under MOC conditions. The goal is to create adaptive, topology-aware frameworks capable of improving fault detection, risk propagation analysis, and operational decision-making for future industrial energy systems.
