The design of rod bundles in nuclear application nowadays is assessed by CFD (computational fluid dynamics). The accuracy of CFD models need validation. Within the OECD/NEA benchmark MATiS-H (Measurement and Analysis of Turbulent Mixing in Sub-channels Horizontal) a single-phase water flow in a 5x5 rod bundle is studied. In the benchmark, two types of spacer grids are tested, the swirl type and the split type, where the current study focuses on the split type spacer grid. Comparison of CFD results obtained at Karlsruhe Institut of Technology (KIT) with experimental results of KAERI (Korea Atomic Energy Research Institute) are presented. In the benchmark velocities components along selected lines downstream of the spacer grid are measured and compared to CFD results. The CFD code STAR CCM+ with the Realized k-E model is used. Comparisons with experimental results show quantitative and qualitative agreement for the averaged values of velocity components. Comparisons of results to other benchmark partners using different modeling show that the selected mesh size and models for the analysis of the current case gives relatively accurate results. However, the used turbulent model (Realized k-E) does not capture the turbulent intensity correctly. Computation shows that the flow has very high mixing due to the spacer grid, which does not decay within the measurements domain (z/ D-H =0-10 downstream of spacer grid). The same conclusion can be drawn from experimental data.
