Investigations of the Application of CFD to Flow Expected in the Lower Plenum of the Prismatic VHTR

The Generation IV (Gen IV) very high temperature reactor (VHTR) will either be a prismatic (block) or pebble bed design. However, a prismatic VHTR reference design, based on the General Atomics Gas Turbine-Modular Helium Reactor (GT-MHR) [General Atomics, 1996] has been developed for preliminary analysis purposes [MacDonald, et al., 2003]. Numerical simulation studies reported herein are based on this reference design. In the lower plenum of the prismatic reference design, the flow will be introduced by dozens of turbulent jets from the core above. The jet flow will encounter rows of columns that support the core. The flow from the core will have to turn ninety degrees and flow toward the exit duct as it passed through the forest of support columns. Due to the radial variation of the power density in the core, the jets will be at various temperatures at the inlet to the lower plenum. This presents some concerns, including that local hot spots may occur in the lower plenum. This may have a deleterious effect on the materials present as well as cause a variation in temperature to be present as the flow enters the power conversion system machinery, which could cause problems with the operation of the machinery. In the past, systems analysis codes have been used to model flow in nuclear reactor systems. It is recognized, however, that such codes are not capable of modeling the local physics of the flow to be able to analyze for local mixing and temperature variations. This has led to the determination that computational fluid dynamic (CFD) codes be used, which are generally regarded as having the capability of accurately simulating local flow physics. Accurate flow modeling involves determining appropriate modeling strategies needed to obtain accurate analyses. These include determining the fineness of the grid needed, the required iterative convergence tolerance, which numerical discretization method to use, and which turbulence model and wall treatment should be employed. It also involves validating the computer code and turbulence model against a series of separate and combined flow phenomena and selecting the data used for the validation. This report describes progress made to identify proper modeling strategies for simulating the lower plenum flow for the task entitled “CFD software validation of jets in crossflow,” which was designed to investigate the issues pertaining to the validation process. The flow phenomenon previously chosen to investigate is flow in a staggered tube bank because it is shown by preliminary simulations to be the location of the highest turbulence intensity in the lower plenum Numerical simulations were previously obtained assuming that the flow is steady. Various turbulence models were employed along with strategies to reduce numerical error to allow appropriate comparisons of the results. It was determined that the sophisticated Reynolds stress model (RSM) provided the best results. It was later determined that the flow is an unsteady flow wherein circulating eddies grow behind the tube and ‘peel off’ alternately from the top and the bottom of the tube. Additional calculations show that the mean velocity is well predicted when the flow is modeled as an unsteady flow. The results for U are clearly superior for the unsteady computations; the unsteady computations for the turbulence stress are similar to those for the steady calculations, showing the same trends. It is clear that strategi