Energy Software S.L. (ENSO) has undertaken an innovative project to address the inherent uncertainties present in the severe accident analysis codes used in nuclear safety assessments. Building on the outcomes of two major international initiatives, the International Atomic Energy Agency Coordinated Research Project I31033 and the HORIZON‑2020 MUSA (Management and Uncertainties of Severe Accidents) project, ENSO has established a solid foundation for identifying and quantifying the main sources of uncertainty in severe accident simulations. These initiatives also revealed key limitations, such as excessive computational time, frequent simulation failures, and truncation effects.The proposed approach follows a stepwise structure, applying the Wilks’ methodology in two consecutive phases: the in-vessel phase, which is focused on accident progression up to reactor pressure vessel failure, and the ex-vessel phase, analyzing the subsequent phenomena occurring within the containment. This stepwise strategy enables the identification and quantification of uncertainty sources at different stages of the accident, providing essential insights for emergency decision making and supporting the design of future research in the field.To validate the methodology, ENSO developed an evaluation model of a Generation-II four-loop Westinghouse pressurized water reactor reactor subjected to a low-pressure station blackout scenario. The RELAP/SCDAPSIM/MOD3.4 code was used for the in-vessel phase, while MELCOR was employed for the ex-vessel containment analysis. An uncertainty analysis was performed, and the early results indicate both the robustness of the proposed methodology and its benefits in terms of computational efficiency and improved calculation stability.However, several challenges emerged during the long-term molten core concrete interaction phase, highlighting the need for further development and refinement.
Bocanegra et al. (Wed,) studied this question.