CFD analysis of pressure drop across grid spacers in rod bundles compared to correlations and heavy liquid metal experimental data

Early studies of the flow in rod bundles with spacer grids suggest that the pressure drop can be decomposed in contributions due to flow area variations by spacer grids and frictional losses along the rods. For these shape and frictional losses simple correlations based on theoretical and experimental data have been proposed. In the OECD benchmark study LACANES it was observed that correlations could well describe the flow behavior of the heavy liquid metal loop including a rod bundle with the exception of the core region, where different experts chose different pressure-loss correlations for the losses due to spacer grids. Here, RANS-CFD simulations provided very good data compared to the experimental data. It was observed that the most commonly applied Rehme correlation underestimated the shape losses. The available correlations relate the pressure drop across a grid spacer to the relative plugging of the spacer i.e. solidity e(max). More sophisticated correlations distinct between spacer grids with round or sharp leading edge shape. The purpose of this study is to (i) show that CFD is suitable to predict pressure drop across spacer grids and (ii) to access the generality of pressure drop correlations. By verification and validation of CFD results against experimental data obtained in KALLA we show (i). The generality (ii) is challenged by considering three cases which yield identical pressure drop in the correlations. First we test the effect of surface roughness, a parameter not present in the correlations. Here we compare a simulation assuming a typical surface roughness representing the experimental situation to a perfectly smooth spacer surface. Second we reverse the flow direction for the spacer grid employed in the experiments which is asymmetric. The flow direction reversal is chosen for convenience, since an asymmetric spacer grid with given blockage ratio, may result in different flow situations depending on flow direction. Obviously blockage develops differently along flow direction due to the asymmetry, so that the same numerical mesh can be used for two distinct simulations. The study shows that prediction of pressure drop across spacer grids by RANS-CFD using well resolved meshes in the spacer grid region yields more accurate results than any available correlation. (C) 2016 Elsevier B.V. All rights reserved.