Page 246 - Thermal Hydraulics Aspects of Liquid Metal Cooled Nuclear Reactors
P. 246
CFD—Introduction 217
6.4 Reduced resolution RANS
Reduced resolution RANS actually means that the geometry is still resolved using
body-fitted grids but that the mesh density is not sufficient according to the best prac-
tice guidelines, for example, published by Casey and Wintergeste (2000). In such
cases, often, the boundary layers near the wall are not well resolved. A more elaborate
description can be found in Roelofs et al. (2012).
In cases where it can’t be afforded to produce a well-resolved mesh, other means of
modeling should be used. A reduced resolution modeling approach does not fulfill the
requirements of a well-resolved CFD simulation, in which the mesh shows mesh-
independent results and in which the boundary layers are taken into account by an offset
layer from the wall ensuring proper wall treatment by aligning the mesh with the wall
with regularly shaped volumes. Basically, a well-resolved simulation should take into
account the best practice guidelines for CFD as reported by Casey and Wintergeste
(2000). Like a well-resolved RANS simulation, a reduced resolution model solves
the well-known equations of conservation of mass, momentum, and energy. However,
as the bulk mesh is relatively coarse and additionally the computational volumes near
thephysical wallsare notwell aligned and regular, the uncertainties insolving theequa-
tions increase. Therefore, prior to application of such a reduced resolution mesh, the
accuracy of the reduced resolution mesh needs to be determined in comparison with
a well-resolved RANS result on a flow configuration that is similar to the actual appli-
cation. This places a reduced resolution approach between a well-resolved RANS sim-
ulation and a low-resolution CFD simulation, as depicted in Fig. 6.1.
6.5 Low-resolution CFD
In cases where it can’t be afforded to produce a well-resolved mesh or alternatively a
reduced resolution mesh, other means of modeling should be used. When applying
CFD, an obvious choice is to switch to a porous medium model or another low-
resolution CFD approach, for example, the coarse-grid CFD approach developed
by Viellieber and Class (2015) or the momentum source model from Hu and
Fanning (2013). When using such a technique, porous media parameters, volumetric
forces, momentum sources, or other modeling parameters need to be applied based on
correlations, well-resolved CFD, and/or experimental data. In fact, this is quite similar
to developments toward three-dimensional nodalization in system thermal-hydraulics
codes as described in Chapter 4.
References
Casey, M., Wintergeste, T., 2000. Best Practice Guidelines. ERCOFTAC Special Interest Group
on ‘Quality and Trust in Industrial CFD’. .
Davidson, P.A., 2004. Turbulence – An Introduction for Scientists and Engineers. Oxford Uni-
versity Press, Oxford.