Modelling of sprays in containment applications with A CMFD code

Abstract

During the course of a hypothetical severe accident in a Pressurized Water Reactor (PWR), spray systems are used in the containment in order to prevent overpressure in case of a steam line break, and to enhance the gas mixing in case of the presence of hydrogen. In the frame of the Severe Accident Research Network (SARNET) of the 6th EC Framework Programme, two tests was produced in the TOSQAN facility in order to study the spray behaviour under severe accident conditions: TOSQAN 101 and TOSQAN 113. The TOSQAN facility is a closed cylindrical vessel. The inner spray system is located on the top of the enclosure on the vertical axis. For the TOSQAN 101 case, an initial pressurization in the vessel is performed with superheated steam up to 2.5 bar. Then, steam injection is stopped and spraying starts simultaneously at a given water temperature (around 25 °C) and water mass flow-rate (around 30 g/s). The depressurization transient starts and continues until the equilibrium phase, which corresponds to the stabilization of the average temperature and pressure of the gaseous mixture inside the vessel. The purpose of the TOSQAN 113 cold spray test is to study helium mixing due to spray activation without heat and mass transfers between gas and droplets. We present in this paper the spray modelling implemented in NEPTUNE_CFD, a three-dimensional multi-fluid code developed especially for nuclear reactor applications. A new model dedicated to the droplet evaporation at the wall is also detailed. Keeping in mind the Best Practice Guidelines, 1 1 Best Practice Guidelines for the use of CFD in Nuclear Reactor Safety Applications, NEA/CSNI/R5 (2007). closure laws have been selected to ensure a grid-dependence as weak as possible. For the TOSQAN 113 case, the time evolution of the helium volume fraction calculated shows that the physical approach described in the paper is able to reproduce the mixing of helium by the spray. The prediction of the transient behaviour should be improved by including in the model corrections based on better understanding of the influence of the dispersed phase on the turbulence of the continuous phase. For the TOSQAN 101 case, droplet velocity, steam volume fraction and gas temperature profiles compare favourably with the experimental results. In the frame of the SARNET network, the results obtained with the physical modelling implemented in the NEPTUNE_CFD code reproduce correctly the entrainment phenomena and the condensation zone ( Malet and Métier, 2007).