In-containment source term predictability of ASTEC-Na: Major insights from data-predictions benchmarking

Modeling the containment response to a sodium pool fire is to be one of the key aspects of any comprehensive safety evaluation of the new generation of sodium cooled fast reactors. Through a peer review of earlier experimental investigations some useful data can be collected and then used for assessing the current analytical capabilities to model severe accidents or some of their specific aspects. This paper provides major insights into the in-containment aerosol behavior predictability of ASTEC-Na (CPA∗ module) during Na-pool fires. By comparing against tests from the ABCOVE (AB1 and AB2) and FAUNA (F2) programs, it has been shown that experimental trends can be roughly reproduced with a single-cell approach whenever natural convection is effective in making the vessel atmosphere uniform both thermally and in composition. Nonetheless, the present heavy parametrization of ASTEC-Na models should be avoided or strongly supported by further experimentation that allows setting sound default values, concerning both combustion energy distribution and aerosol formation and distribution. Anyway, the peer data review has highlighted that a meaningful comparison to predictions is not always feasible due to large data uncertainties, particularly at the beginning of Na burning. As for the particle ageing, the comparisons set seems to indicate that transformation from oxides to hydroxides is predicted to be too slow; nevertheless, a more extensive benchmarking should be conducted to confirm it.