地球シミュレータセンター

平成19年度 地球シミュレータ利用報告会

1. プロジェクト名

Direct numerical simulation of turbulent sodium flows in subchannels of an LMFBR fuel subassembly

2. プロジェクト責任者名

Hisashi Ninokata (Tokyo Institute of Technology)

3. プロジェクトの目的

A major objective of this research is to establish a computational simulation-based design and safety approach in nuclear engineering, in particular in the area of nuclear fuel pin subassembly thermal hydraulics design and safety. Emphasis is placed on the delineation, in-depth understanding and modeling of the complex turbulent flow structure inside nuclear fuel pin subassemblies characterized by non-homogeneous and anisotropic turbulence. Applications of Direct Numerical Simulation of turbulence and Large Eddy Simulation techniques are fully employed being aimed at providing higher quality of computational data that should be equivalent to or more detailed information than those provided by experimentation. Also the computational results will be used for engineering modeling as well as the basis on which design data base is constructed.

4. 今年度当初の計画

Calculations should be carried out for the eccentric annulus channel flows to confirm the fully developed flow condition and flow stabilization, and then establish turbulence flow data base that should be useful for engineering applications, in particular in connection to the turbulent flows in tight lattice nuclear fuel pin subassemblies. The targeted phenomena include the local laminarization and global pulsation phenomena. There, influences of the anisotropic turbulence structure and eddy migration behaviors in the non-uniform flow channels will be investigated in detail.

5. 今年度得られた成果、および達成度

成果

Several DNS computations have been performed for the concentric and eccentric channels. The data collected has been used to evaluate different SGS model in order to develop an effective LES methodology in boundary fitted coordinates. An a priori analysis as well as an a posteriori analysis of the flow in concentric annuli and eccentric annuli has been carried out. Several models have been tested, among which, the dynamic Smagorinsky, the dynamic mixed model, the self-similarity model and another variants of the dynamic model ([1]).

The dynamic model and its variant performed fairly well from the point of view of a priori and a posteriori tests. They may be considered the ideal choice for the simulation of the flow in annular channels and rod-bundles.

Then, an extensive LES computational campaign has been performed for the eccentric channel at different values of the Reynolds number (Figure 1) and the eccentricity to investigate the characteristics of the flow in eccentric channels.

Important aspects of the flow field in concentric and eccentric annuli have been confirmed and reproduced through the present methodology ([2]). In particular, the effect of transverse curvature on the inner wall, as well as the effect of eccentricity on the wall shear stress, has been successfully simulated. Moreover, some fundamental observations have been possible:

• 1. flows in eccentric annuli are characterized by a strong Reynolds effect in the narrow gap in the Reynolds region surveyed;
• 2. flows in eccentric annuli present secondary flows, whose shape is constant in the range of Reynolds number surveyed but depends upon the value of eccentricity;
• 3. at low Reynolds numbers the secondary flows are associated to coherent periodic structures in the direction of homogeneous turbulence, the flow is characterized by high coherence lengths;
• 4. the condition of anisotropy of turbulence in the narrow gap changes greatly as the Reynolds number increases .

From previous works and the evidence presented here it appears that the transition to turbulence in geometry such as the eccentric annuli is accompanied by the formation of a street of counter-rotating vortices in the region near the narrow gap. These coherent structures (Figure 2) persist at low Reynolds numbers but they progressively become less dominant, at least for an eccentricity equal to 0.5, as the Reynolds number increases. Contemporarily, in the narrow gap the local profile of the streamwise velocity evolves from a purely laminar solution to a solution characterized by the presence of turbulence production near walls. The shear stress in the narrow gap region evolves from an almost laminar condition for a Reynolds number equal to 3200 to an increasingly turbulent solution.

At low Reynolds number and eccentricit equal to 0.5 the relative dominance of the coherent structures is associated with a strong anisotropy in the narrow gap (turbulence has a local two-component pattern), while at higher Reynolds numbers a nearly isotropic condition is recovered far from the walls.

At higher values of eccentricity (0.8) the structures persist even at Reynolds as high as 11300, and anisotropy continues to be dominant in the narrow gap. It appears that transition to fully developed turbulence in the narrow gap is delayed at eccentricity equal to 0.8 if compared to an eccentricity equal to 0.5. In general the delay in transition between the wide gap and the narrow gap depends upon the eccentricity.

In conclusion, evidence based on this and previous research suggests that in eccentric channels coherent structures are present for some combination of the geometric parameters and Reynolds number. Anisotropy conditions in the narrow gap also depend upon the value of the geometric parameters and Reynolds number. In particular, higher value of eccentricity imply that a strong anisotropy persist in the narrow gap even at higher Reynolds numbers (as high as 11300 for eccentricity of 0.8). A POD study of the fields obtained in this study have been performed to obtain additional insight in the physics of flow in eccentric channels and other geometries ([3]).

達成度

100%