Swirling spacer grids in a model of nuclear fuel assembly - PIV data of isothermal flow
Description
The swirl grid produces longitudinal vortices , which decay much slowlier than general grid-generated turbulence. This effect has been demonstrated for square lattice of such generated vortices without any fuel rods, see the article https://doi.org/10.1063/5.0207124. The idea of this experiment is to apply these effect to the case of nuclear fuel assembly, where are used special mixing grids enhancing mixing of hot coolant at the boundary layers of the fuel rods with the cooler liquid in the inter-rod channel centers. However, this mixing efficiency decreases fast, while the longitudinal vortices shift the decay far downstream. The database contains PIV vector fields of the wind tunnel test of isothermal flow through a model of nuclear fuel assembly. The model is upscaled, hence the diameter of the fuel rods is 42.2 mm (real is 9.1 mm), the distance of fuel rods (the mesh parameter M) is 59.1 mm (real is 12.75 mm). The fuel rods are physically modeled by using water pipes of length 3 m, 5 spacer grids are placed in total. The following configurations are tested: - all grids just base spacer grids (name e.g.: "U20ms-base-z02") - the last grid before the PIV plane contains the swirling elements of Rossby number 2.5 (name e.g.: "U20ms-1vir4-z02") - the last two grids are the swirler grids with oppositely oriented swirling elements (name e.g.: "U20ms-2vir4-z02") - only single swirl grid without fuel rods and without any other grids (name e.g.: "U20ms-1pv4-z02") All test are performed with atmospheric air at velocity 20 m/s. The last number in the name ("...-z04") is the distance from the PIV plane to the last grid in multiples of M; there are 5 distances: 0.5, 1, 2, 3 and 4. The PIV plane is perpendicular to the rod axis. PIV is in stereo configuration with two cameras. If some author is interested to process our image data by his/her own method, we are open to share the original images (638 GB), contact us.
Files
Steps to reproduce
Grid: 3D printed by using Prusa MK4 from PLA. Due to size, the grid is printed in three pieces glued together. In the CAD for printing use sestava("S"); PlacickyS();", then sestava("L"); PlacickyL(); and sestava("P"); PlacickyP(); The model will contain small plates to increase the adhesion to printing plate. The connection between grid parts is done by thinner walls, which can be glued. We used Bison Plastic Adhesive, no. 87296105. Fuel rods: as fuel rods we used plastic pipes from company FV - plast a.s. Four tubes are cutted along axis in half in order to fit hexagonal symmetry into rectangular test section. PIV data processing: Time between laser pulses is 40 μs. Repeating frequency of image pairs is 11 Hz, therefore no temporal development can be observed and the snapshots can be considered as statistically independent. Note that the PIV original data are the pair of 12-bit depth images for each snapshot, therefore they are quite large, here we publish only the vector fields calculated by the Dantec Dynamic studio software (function AdaptivePIV, grid step: 32 pix, min IA: 16 pix, max IA: 32 pix, no Window/Filter functions, validation by Universal Outlier Detection on 5×5 neighborhood, min. normalization 0.1, acceptance limit 2.0, IA adapted to velocity gradients with |Ux| < 0.1, (Ux^2)^0.5 < 0.2 in max. 10 iterations).
Institutions
- Zapadoceska univerzita v Plzni Fakulta strojni