CFD simulations of helium layer erosion in one PANDA vessel following cooler activation

Abstract

We performed CFD simulations of the two H2P6 cooler tests conducted in the PSI PANDA vessel as part of the OECD/NEA HYMERES-2 project. Initially, the vessel atmosphere consists of a 60/40 % vol. steam- air mixture. Then, a helium stratification (15 % vol.) is introduced in the upper half of the vessel. The two experiments H2P6_1 and H2P6_2 are characterized by the activation respectively of 3 and 1 cooler(s) in the upper part of the vessel. In the course of the experiments, condensation takes place, resulting in a gradual pressure decay and erosion of the helium layer until complete mixing is achieved.The URANS k-ε is used to model turbulence. CFD Best Practice Guidelines are followed to minimize numerical errors, with preliminary mesh sensitivity studies. A correlation based on the heat-and-mass transfer analogy is employed to estimate steam condensation rates.Each cooler consisting of 18 tubes is simplified into a cuboid with equivalent heat transfer area. In a further simplification, it is assumed that the cooler wall has a thickness of zero and is initially at the temperature of the coolant. The small initial energy stored in the mass of the cooler at time zero is therefore neglected.The simulation for H2P6_1 matches the main experimental parameters (heat removal rate, pressure, species concentrations) almost perfectly. For H2P6_2 the heat removal rate is slightly under-predicted, with a corresponding over-prediction of the pressure by a maximum of 3 %. The inclusion of radiative heat transfer modeling improves the accuracy of the temperature field but has little influence on the pressure or heat removal histories.These results show that CFD allows accurate predictions of containment mixing when coolers are activated. Moreover, the simulations also indicate that multiple cooler units can be represented by one larger equivalent unit.