Liquid spray modeling under sodium fire accidents

Abstract

The Sodium-Cooled Fast Reactor (SFR) system, which features a fast-spectrum, sodium-cooled reactor and a closed fuel cycle for efficient management of actinides and conversion of fertile uranium, is rather prominent in the nuclear reactor design in Generation IV (Gen IV), and likely to be widespreadly used and well-constructed all over the world. However, molten sodium may be sprayed into the ambient or secondary side of steam generator, induced by the pressure drop, in case of an inevitable break in the primary system. This kind of leakage can result in sodium spray fire accident, as a result of the chemical activity of sodium, significantly damaging the equipment and working staffs nearby. Sodium spray fire, which occurs not only in the nuclear industry, but also in solar industry, is a complex process consisting of several specific phenomena, such as the spray dynamics (e.g. droplet particle size/velocity distribution, particle collision and agglomeration), droplets evaporation, sodium aerosol diffusion, thus being drawn many experts’ attention. Based on International Atomic Energy Agency (IAEA)’s effort on phenomena identification and ranking table (PIRT) for sodium spray fire, prediction of droplet size distribution and the mean diameter in particular, are supposed to be in the very first place at level 2, due to the lack of available experimental data for both of model development and validation, and code inadequacy, as well. In this paper, sodium droplet size distribution model is theoretically derived, regarding both of mass and momentum equations as constraints, based on the maximum entropy principle, which can quantitatively describe the effects of two main influencing factors, including initial injection velocity and thermophysical properties of molten sodium. This analytical model is implemented in three typical conditions, including two sodium droplets size distribution experiments and one water droplet experiment. Validations against both of the available test data and predictions by an empirical model, named of Nukiyama-Tanasawa correlation, are conducted with the relative error of median diameter ranging from 17.3% to 18.2%, which ensures the reliability and feasibility of the new model. Moreover, sensitivity analyses are carried out, in terms of typical test conditions of sodium spray fire, which demonstrates that the median diameter of droplets’ size decreases with both of the increase of initial injection velocity and sodium temperature, thus probable triggering positive effect on the propagation of the accident, due to rapid interaction of violent chemical reaction and sodium spray. Besides, the present work can provide a reliable analytical model for the development of safety analysis codes for SFRs.