Capabilities of high y + wall approaches in predicting heat transfer to supercritical fluids in rod bundle geometries

The present paper summarises the results of the simulation of heat transfer to supercritical water in rod-bundle geometries by a CFD code, using wall function models. Two different sets of experimental data were considered, concerning both relatively high and low mass flux conditions and inlet temperature spanning from low to near-critical values. In past analyses, the unsuitability of low-Reynolds number turbulence models was observed in predicting heat transfer in rod bundles, when both high mass and heat flux values were imposed; in fact, large overestimations of wall temperatures were reported in such conditions. This motivated to try simpler models, such as the wall function approach, in order to investigate if, though expectedly not very accurate, they could at least reproduce the experimental data at an acceptable level. As reported in the present paper, the selected models, which adopt a “high y+” wall treatment (indicating wall functions in the STAR-CCM+ code), seem reasonably able at reproducing the general observed experimental trends. The present understanding of the phenomena and the available modelling techniques unfortunately do not allow obtaining better results when the wall or fluid temperature approach the pseudocritical value. However, the comparison with results obtained by low-Reynolds models shows that, at least when facing operating conditions similar to the ones considered in the present work, a downgrading of the adopted modelling techniques may be beneficial and allows obtaining reasonable results. Analyses were also performed considering conditions far from the pseudo-critical temperature, in which low-Reynolds models had provided good performance, reporting that the wall function approach seems effective also in these cases for obtaining first guess results.