Further development and validation of OpenFOAM solvers and coupling methods for thermal hydraulics in the cooling cycle (WEVAFOAM). Final report

Within the BMUV- and previously BMWK-funded project RS1591 "Further development and validation of OpenFOAM solvers and coupling methods for thermal hydraulics in the cooling circuit" (WEVAFOAM), the open-source code OpenFOAM was validated for the analysis of complex single-phase flows in PWR cores and the existing ATHLETOpenFOAM coupling interface for a selected two-phase flow pattern was further developed and validated on the two-phase LSTF ROSA V Test 1.1 experiment. Today, fuel assembly spacer grids with complex geometries are used to improve the heat transfer in fuel assemblies through increased flow turbulence. The CFD program OpenFOAM has to be validated to reliably predict the locally occurring flow and heat transfer phenomena in the fuel assemblies. For this purpose, analyses were carried out of the Korean MATiS-H Benchmark for fuel assembly configurations with two different types of spacer grids. RANS, URANS and the scale-resolving simulations with Open- FOAM and ANSYS CFX showed that similar results could be achieved with both programs, whereas the calculated velocity profiles agree well with the measured data. The flow investigations in the PWR cores were continued within the GRS participation in the AER VVER-440 benchmark. The benchmark deals with coolant mixing and temperature distribution in two VVER-440 fuel assembly types. These have 11 spacers and different heating. The CFD results achieved with OpenFOAM and ANSYS CFX agree very well. Although the benchmark is not yet complete, GRS has already sent the generated numerical data to the coordinator. Due to the growing importance of the coupled ATHLET-CFD analyses, ATHLET Open- FOAM was validated against the two-phase ROSA V Test 1.1 experiment in the LSTF facility. First, though, the existing coupling had to be further improved. New boundary conditions that allow the calculation of two-phase, free-surface flow have been programmed. In the two-phase coupled analyses, the LSTF cold leg A and half of the downcomer were represented with OpenFOAM, while the rest of the Japanese facility was modelled with ATHLET. The complex analyses were carried out with a very small-time step, so that numerical instabilities could be mitigated with interface adjustments. The results achieved with ATHLET-OpenFOAM agree well with experimental data for most measurement locations. However, there are positions for which the fluid temperature is overestimated or underestimated by the computer program. This is because of deficiencies in the CFD calculation and not the interface, since the stand-alone Open-FOAM and the ATHLET-OpenFOAM results agree very well.