CFD Model for a Jet">
Nuclear Technology / Volume 171 / Number 2 / August 2010 / Pages 161-170
Technical Paper / Thermal Hydraulics
This work is a three-dimensional (3-D) implementation of the computational fluid dynamics (CFD) model for a shutdown boron injection jet of a pressurized heavy water reactor, previously developed for the axisymmetric case. The boron shutdown system injects round boron jets into a moderator tank with an array of cylindrical coolant channels. The boron injection jets are tilted with respect to the coolant channels. The 3-D formulation allows the calculation of the curved trajectory of a jet that is deflected by the coolant channels. Furthermore, the modeling of the turbulent jet mixing is performed with a realizable k- model to obtain the concentration of boron around the jet axis. The final objective is to predict the distribution of boron inside the moderator tank to calculate the insertion of negative reactivity into the reactor during a fast shutdown with a multidimensional PARCS/RELAP5 coupled model. The implementation of the present CFD results into PARCS/RELAP5 and the neutronic results are discussed in a separate paper.
A porous-medium approach is used to represent the coolant channels. This porous-medium methodology is based on a volume average of the governing equations that is equivalent to the two-fluid model used for two-phase flows. The additional source terms that appear because of the averaging (i.e., constitutive relations) in the present model are related to drag over an array of cylinders (i.e., the fuel channels) for the momentum equation and additional mixing source terms due to the cylinders for both the turbulent kinetic energy and the turbulent dissipation transport equations.
The CFD model is validated with experimental data of the boron concentration distribution obtained in a 1:7.66 scale facility representing the jets and the moderator tank. Good agreement is achieved for the trajectory of the jet centerline. The transverse spreading of the boron due to turbulence is also well predicted, though the CFD results somewhat overpredict the peak concentration compared with the measurements.