As nuclear technologies evolve, risk assessment methodologies must adapt to account for new types of uncertainties. In this context the NEA held the International Workshop on the Treatment of Uncertainties for Novel Aspects of Risk Analyses from 31 March to 2 April 2025 in Madrid, Spain.
The event, organised by the NEA Working Group on Risk Assessment (WGRISK) and hosted by the Spanish Nuclear Safety Council (CSN), brought together experts from around the world to discuss state-of-the-art approaches to handling uncertainties in risk analysis. The event facilitated high-level discussions on risk assessment methodologies and their application in diverse domains and featured contributions from distinguished speakers.
CSN’s president, Juan Carlos Lentijo, delivered a welcome message that set the stage for the discussions, while Prof. George Apostolakis of the Central Research Institute of Electric Power Industry (CRIEPI) gave a keynote speech on uncertainties in risk-informed decision making, providing insight into the complexities of uncertainty management in modern risk assessments.
Prof. George Apostolakis (CRIEPI) delivers keynote remarks.
While traditional probabilistic safety assessments (PSAs) are well-established for conventional reactor designs, novel aspects such as passive safety systems, digital instrumentation and control, and multi-unit dependencies introduce additional complexities. Uncertainty treatment is a critical factor in ensuring robust risk-informed decision making.
The workshop aimed to:
- identify key sources of uncertainty in novel risk assessment scenarios;
- discuss best practices and methodologies for quantifying and managing uncertainties;
- explore the role of dynamic probabilistic safety assessment (DPSA) and other advanced tools in reducing uncertainty; and
- facilitate international collaboration on uncertainty treatment approaches.
The workshop covered several key topics, including the classification and management of uncertainties through the comparison of existing methods and emerging approaches and by discussing practical applications for regulators and industry. Participants also exchanged views on DPSA; enhancing risk assessments through machine learning and artificial intelligence (AI); and frameworks for risk-informed decision making by integrating uncertainty considerations into organisational decision processes and fostering collaboration between regulators and industry.
Despite advancements in probabilistic risk assessment methodologies, several challenges remain:
- Standardisation and harmonisation: There is a need for common frameworks and methodologies to ensure consistency in uncertainty treatment across different regulatory bodies and industrial applications.
- Validation of advanced models: It is important to ensure the reliability of AI-based and dynamic probabilistic approaches in risk assessments.
- Integration into decision making: Uncertainty management must be incorporated effectively into regulatory frameworks and operational risk management strategies.
- Addressing data limitations: Improvements are needed in data collection, sharing and quality to support more accurate uncertainty quantification.
- Communication: Uncertainty information should be provided in an understandable way to stakeholders.
The discussions offered new insights into handling uncertainties and provided recommendations for future research and regulatory developments. The workshop participants emphasised the importance of continued research and international co-operation in tackling these challenges.
