Modelagem numérica com validacao experimental de escoamento através do bocal inferior de um elemento combustível nuclear

Abstract

The commercial CFD program, ANSYS CFX 10.0, was used to define a numerical procedure to evaluate the pressure drop in a permanent, incompressible and adiabatic flow through the standard Bottom end Piece (BP) of a nuclear fuel assembly. The numerical results were compared to experimental pressure loss results obtained at the Reactor Hydrodynamic Circuit (CHR) of the Centro Tecnológico da Marinha (CTMSP). Initial studies on 1/8 symmetry perforated plates, which are the predominant geometry of the BP, were performed to define the optimal mesh parameters and adequate turbulence model. The RANS two equations turbulence models k-e, k-. and SST were assessed. The influences of the chamfers of the BPs perforated plate orifices, of a gap between the BP and the duct present in the experiments, and of the geometry upstream of the test section were also appraised. The simulations confirmed that the presence of chamfers causes a drastic reduction of both pressure loss (~40%) and pressure recuperation length. The gap causes a reduction of ~5% of the pressure loss due toboth the flow near the wall and the increase of the cross section area of the duct. The increase in the mesh prevented, due to computational limitations, the simulation of the gap with the BP, but a compensation of the mass flow due to the increase of the cross section area was applied. It was observed that the geometry upstream of the test section promotes a nonuniform velocity and turbulence profiles that cause a ~4% reduction of the pressure loss at the perforated plate. The presence of this upstream geometry was omitted in the final BP simulation model for it disables the use of a 1/8 symmetry, which would prevent the simulation. Then a 1/8 symmetry of the BP was simulated with the turbulence models k-e and SST for all experimental conditions, applying a uniform profile of velocity and turbulence at the inlet. The pressure loss results obtained with the SST and k-e models were in average ~19% and ~10% higher, respectively, than the experiments. The discrepancies observed can be mainly associated to the simplifications, performed due to computational limitations, that shown a tendency to reduce the calculated pressure loss. This work shall support the development of a new advanced nuclear fuel element at INB.