Research
My research develops rigorous and practical methods for advancing probabilistic risk assessment (PRA) in operating and advanced nuclear power plants. I focus on how uncertainties, human actions, physical processes, and digital technologies interact over time and shape system risk. Across this work, I integrate human reliability analysis, advanced modeling and simulation, AI-driven automation, digital twins, and virtual reality to support safety, decisionmaking, and regulation in complex socio-technical systems.
Browse linked publications for each theme, or reach out for collaboration.
Theme 1: Advancing PRA for Operating and Advanced Nuclear Power Plants
Problem: PRA models must support both near-term operating decisions and long-horizon design and licensing decisions for advanced reactors, while retaining technical rigor under changing plant conditions.
Methods and contributions: This work develops spatiotemporal PRA methods, probabilistic validation workflows, and risk-importance ranking approaches that improve model realism and traceability for risk-informed analysis.
Representative applications: Fire PRA realism enhancement, uncertainty-based maintenance-degradation coupling, and decision support for technology deployment and operational flexibility.
Theme 2: Spatiotemporal Modeling of Coupled Human-Physics-Digital Technologies
Problem: Safety-critical decisions are influenced by coupled interactions among human performance, physical failure mechanisms, and digital systems, but these dependencies are often simplified or separated in practice.
Methods and contributions: The research integrates human reliability analysis, physical phenomena simulation, with physics-informed and AI-driven digital systems (e.g., automation, digital twins) to represent temporal dependencies, organizational factors, and automation interactions in a unified socio-technical framework.
Representative applications: External control room emergency response analysis, human-digital twin interaction modeling, and integrated treatment of coupled human-technology effects in PRA.
Theme 3: Validation and Uncertainty Quantification with Sparse Data
Problem: Validation or trustworthiness evaluation for high-consequence and novel systems frequently faces sparse, uncertain, or heterogeneous data, creating gaps between model outputs and defensible decisions.
Methods and contributions: This theme advances probabilistic validation theory and computational platforms to quantify epistemic and aleatory uncertainty, evaluate model validity and trustworthiness, and support transparent technical interpretation.
Representative applications: Validation of fire progression models in Fire PRA and coupled maintenance-degradation models under data limitations, as well as AI-driven automation trustworthiness assessments.
Theme 4: Risk-Informed Decisionmaking and Regulation
Problem: Technology-inclusive and performance-based regulatory pathways require methods that can bridge technical and legal uncertainties, operational realities, and policy implementation.
Methods and contributions: The research develops context-based risk-informed analysis, regulation-supporting validation approaches, and decision frameworks for integrating evidence from PRA, simulation, and socio-technical studies.
Representative applications: NRC-aligned studies on advanced reactor regulation, OECD-NEA NEST collaborations, and contributions to risk-informed approaches for deployment and oversight.