High-thermal performance PWR (pressurized water reactor) spacer grids require both low pressure loss and high critical heat flux (CHF) properties. Numerical investigations on the effect of angles and position of mixing vanes and to understand in more details the main physical phenomena (wall boiling, entrainment of bubbles in the wakes, recondensation) are required.In the field of fuel assembly analysis or design by means of CFD codes, the overwhelming majority of the studies are carried out using two-equation eddy viscosity models (EVM), especially the standard K-epsilon model, while the use of Reynolds Stress Transport Models (RSTM) remains exceptional.But extensive testing and application over the past three decades have revealed a number of shortcomings and deficiencies in eddy viscosity models. In fact, the K-epsilon model is totally blind to rotation effects and the swirling flows can be regarded as a special case of fluid rotation. This aspect is crucial for the simulation of a hot channel in a fuel assembly. In fact, the mixing vanes of the spacer grids generate a swirl in the coolant water, to enhance the heat transfer from the rods to the coolant in the hot channels and to limit boiling.First, we started to evaluate computational fluid dynamics results against the AGATE-mixing experiment: single-phase liquid water tests, with Laser-Doppler liquid velocity measurements upstream and downstream of mixing blades. The comparison of computed and experimental azimuthal (circular component in a horizontal plane) liquid velocity downstream of a mixing vane for the AGATE-mixing test shows that the rotating flow is qualitatively well reproduced by CFD calculations but azimuthal liquid velocity is underestimated with the K-epsilon model.Before comparing performance of EVM and RSTM models on fuel assembly geometry, we performed calculations with a simpler geometry, the ASU-annular channel case. A wall function model dedicated to boiling flows is also proposed. (C) 2009 Elsevier B.V. All rights reserved.

High-thermal performance PWR (pressurized water reactor) spacer grids require both low-pressure loss and high critical heat flux (CHF) properties. Numerical investigations on the effect of angles and position of mixing vanes and to understand in more details the main physical phenomena (wall boiling, entrainment of bubbles in the wakes, recondensation) are required.In the field of fuel assembly analysis or design by means of CFD codes, the overwhelming majority of the studies are carried out using two-equation Eddy Viscosity Models (EVM), especially the standard K-e model, while the use of Reynolds Stress Transport Models (RSTM) remains exceptional.The simulation of swirling flow generated by the mixing vanes plays an important role for the prediction of the CHF for the fuel assemblies. For this reason, according to Mimouni et al. (2008b, 2009b), rotation effects and RSTM model are more specifically addressed in the paper.Before comparing performance of EVM and RSTM models on fuel assembly geometry, we performed calculations with simpler geometries, the DEBORA case and the ASU-annular channel case. ASU-annular channel case has already been addressed in Mimouni et al. (2008b, 2009b).Then, a geometry closer to actual fuel assemblies is considered. It consists of a rectangular test section in which a 2 x 2 rod bundle equipped with a simple spacer grid with mixing vanes is inserted. The influence of the turbulence model on target variables linked to CHF limitation will be discussed. Moreover, the sensitivity to the mesh refinement will be particularly examined. The study of this case is a further step towards the modelling of the two-phase boiling flow in real-life grids and rod bundles. (C) 2009 Elsevier B.V. All rights reserved.

In our work in 2008, we evaluated the aptitude of the code Neptune_CFD to reproduce the incidence of a structure topped by vanes on a boiling layer, within the framework of the Neptune project. The objective was to reproduce the main effects of the spacer grids. The turbulence of the liquid phase was modeled by a first-order K-ε model. We show in this paper that this model is unable to describe the turbulence of rotating flows, in accordance with the theory. The objective of this paper is to improve the turbulence modeling of the liquid phase by a second turbulence model based on a Rij-ε approach. Results obtained on typical single-phase cases highlight the improvement of the prediction for all computed values. We tested the turbulence model Rij-ε implemented in the code versus typical adiabatic two-phase flow experiments. We check that the simulations with the Reynolds stress transport model (RSTM) give satisfactory results in a simple geometry as compared to a K-ε model: this point is crucial before calculating rod bundle geometries where the K-ε model may fail.

Two pressure-correction algorithms are studied and compared to an approximate Godunov scheme on unsteady compressible cases. The first pressure-correction algorithm sequentially solves the equations for momentum, mass and enthalpy, with sub-iterations which ensure conservativity. The algorithm also conserves the total enthalpy along a streamline, in a steady flow. The second pressure-correction algorithm sequentially solves the equations for mass, momentum and energy without sub-iteration. This scheme is conservative and ensures the discrete positivity of the density. Total enthalpy is conserved along a streamline, in a steady flow. It is numerically verified that both pressure-correction algorithms converge towards the exact solution of Riemann problems, including shock waves, rarefaction waves and contact discontinuities. To achieve this, conservativity is compulsory. The two pressure-correction algorithms and the approximate Godunov scheme are finally compared on cases with heat source terms: all schemes converge towards the same solution as the mesh is refined.

Security in nuclear power plants demands severe limitations of the maximal drop time of rod cluster control assemblies. In February 1995, several assemblies of the Chinese plant in Daya Bay failed to comply with these requirements. Electricité De France undertook a research program to get a better insight of this problem since the plant has been built by French and also because the French new four-loops N4 reactor was equipped with the same guide tubes. This paper is limited to a numerical study of the influence of the pressure forces applied to control rods and due to flow circulation through the guide tubes. After a validation test case, a first calculation has been carried out on a simplified N4 guide tube. The sensitivity of the pressure forces to transverse flow and to modifications of the geometry has been determined. The program has been extended to guide tubes used in 1300-MW reactors and similar computations have been done. To make simulations more representative, a global computation of the flow in the whole upper internals plenum (UIP) will be achieved to provide accurate boundary conditions for local calculations with better resolution.

CFD applications in the context of power plants illustrate the detailed three-dimensional studies in complex geometries that are possible today, with a special emphasis on the necessity of solving the time-dependent equations. These routine simulations give a much more detailed representation of many flow configurations than was possible with scale models in the design stage of the plants, thus increasing their safety and life span. Aside from these global simulations conducted with standard models, use of Second Moment Closures is shown to better represent the effects of body forces, such as buoyancy and rotation, encountered in sub components of the plant. Finally Large Eddy Simulation is introduced on unstructured meshes and the related needs concerning fluid/structure coupling and unsteady thermal loading in plant components are described.

The benefit of using unstructured meshes for industrial applications of Computational Fluid Dynamics is now well established. The main advantages of unstructured meshes are to enable an easy and fast (with the use of automatic mesh generator) treatment of complex boundaries and to allow local and automatic mesh refinement. Finite Element methods provide a rigorous mathematics formulation. Nevertheless, implementing complex modelization is more tricky than when using finite volume or finite difference methods on structured meshes, especially when dealing with complex sets of non linear equations such as those encountered in turbulence modelling. This paper presents recent works done at the Research and Development Division of Electricite de France on the implementation of various kinds of turbulence models in the finite element code N3S. First will be detailed the numerical procedure used for solving efficiently the Reynolds Averaged Navier-Stokes Equations (RANSE) coupled with a standard k-epsilon turbulence model. Then, improvement of the k-epsilon model will be presented. The extension of this algorithm to more complex turbulence models such as Reynolds Stress Transport Models (RSTM) will be discussed. The last part of the paper will be devoted to the discussion of Large Eddy Simulation (LES) on unstructured meshes.