Assessment of Uncertainties Effect on Accident Progression and Fission Product Release in a Spent Fuel Pool

Numerical tools are widely used in the nuclear community to assess the behavior of Nuclear Power Plants (NPPs) during postulated accidents, including Severe Accidents (SAs). After the events occurred in Fukushima in 2011, the vulnerability of nuclear fuel stored in Spent Fuel Pools (SFPs) was highlighted and the reliability of the SA codes predictions in SFPs questioned. In this context, the Work Package 6 (WP6) of the H2020 Management and Uncertainties of Severe Accidents (MUSA) project was designed to gain insights on SA codes capabilities for SFP accidents and, more specifically, to know about the uncertainties associated to key output variables. The present study aims at quantifying the uncertainties affecting a SFP analysis, with a focus on accident progression and Fission Products (FPs) release. In particular, Uncertainty Quantification (UQ) methodologies and sensitivity analysis techniques have been applied to the computation of a Fukushima-like scenario in a SFP. The scenario was decided to be a loss-of-cooling scenario without water injection recovery with the calculation starting at the onset of fuel uncover and ending when the amount of fuel lost from initially intact components and transferred to degraded components reaches 1%. Calculations were performed with the SA MELCOR 2.2 code (version 2.2) complemented with the DAKOTA tool for the statistical analysis. For the Uncertainty Analysis (UA), uncertain parameters relative to fuel assemblies’ behavior and degradation and to FPs release and transport were selected. A subsequent sensitivity analysis, based on Pearson’s and Spearman’s Correlation Coefficients (CCs), has also been performed with the intent to highlight the most influencing parameters on the selected Figures Of Merit (FOMs), being these the onset time of FPs release from fuel and the total release of main FPs from fuel as mass fraction of the initial inventory (Xe, Cs, Ru). Outcomes from the UA shows that time evolution of the FOMs related to FPs release follows the same behavior over time. Moreover, uncertainties linked to the same FOMs spread over a narrow band with the BE case results within the uncertainty band. In addition, some best practice and lessons learned can be extracted from this study, i.e., the importance of establishing an accurate reference case prior to the UA, the consistency screening of individual cases, and the lack of a “silver bullet” to conduct the sensitivity analysis.