High-fidelity PIV measurement of cross flow in 5 × 5 rod bundle with mixing vane grids

Abstract

In order to achieve high-fidelity PIV measurement for cross flow in rod bundles, the measurement approach with the two-dimensional (2D) PIV has been optimized systematically and the systematic and random errors are analyzed. The optimizations, covering the design of test section, the setting of optical system, the PIV parameter configuration and the PIV correlation algorithm, were established based on massive sensitivity analysis. Utilizing the optimized PIV configuration, cross flow measurements were conducted in a 5 × 5 rod bundle with the rods 9.5 mm in diameter and the rod pitch of 12.6 mm, which is the typical configuration of fuel assembly in pressurized water reactors (PWRs). The FEP tubes filled with deionized water are employed as the matched index of refraction (MIR) rods to realize uniform illumination across the rod bundle. On each investigated cross section the cross flow was measured in the 16 inner subchannels. In order to maximize the magnification factor, four subchannels are measured in a single test run. Several substantial error sources were identified and accounted for in the error and uncertainty analysis. The systematic error consists mainly of the perspective error and displacement variation error, which were quantified and used to correct the measured results. The uncertainty of RMS velocity is the combination of random error and error due to limited sampling number, while the uncertainty of mean velocity equals to the error resulted from limited sampling number assuming that the systematic error is perfectly compensated. In the experiment the relative standard uncertainty of RMS velocity (to the bulk velocity) is within 0.6%, and the relative standard uncertainty of mean velocity (to the bulk velocity) is less than 0.4%. The measurements were carried out with the subchannel Reynolds number (Re) of 39600 at five cross sections, i.e. 3Dh, 5Dh, 10Dh, 15Dh and 20Dh downstream of a spacer grid. The experimental results showed the detailed cross flow features, including the evolution of vortices, in the rod bundle. The measured lateral flow was substantially dominated by the mixing vanes. Drastic turbulence generation by mixing vanes was observed downstream of a spacer grid. The generated turbulence decayed rapidly within 10Dh downstream of the spacer grid and then damped gradually.