Project Highlights
Making Safe, Clean Nuclear Energy Available Globally
 |
This image represents the turbulent flow of coolant into a mock-up of the upper plenum of an advanced recycling nuclear reactor. The colors indicate the speed of the fluid, red representing regions of high velocity, and blue representing regions of low velocity. Coolant enters from hexagonal channels at the bottom of the plenum as two jets, each with a mean flow velocity of 1 meter-per-second, and exits from single rectangular channel at the top. The results of these simulations will be used in conjunction with data from Argonne's MAX experimental test stand to better understand the thermal-hydraulic properties of the flow. This simulation was performed using the Nek5000 code employing 68826 spectral elements of order N=7 and run on 8192 core of the IBM Blue Gene/P at the Argonne Leadership Computing Facility (ALCF) at Argonne National Laboratory. The ALCF is supported by the Office of Science of the U.S. Department of Energy under contract DE-AC02-06CH11357. [watch video][high-res] |
As part of the nuclear energy program, the United States is committed to new technologies that will dramatically expand the availability of safe, clean nuclear energy to help meet the growing global energy demand. Liquid-metal-cooled fast reactors are a key component of this strategy in that they permit recycling of nuclear fuel and are expected to be economical sources of power.
Approach
Through U.S. Department of Energy (DOE) INCITE allocations, researchers are carrying out large-scale numerical simulations of turbulent thermal transport in sodium-cooled reactor cores. These simulations will enable researchers to gain an understanding of the fundamental thermal mixing phenomena within advanced recycling reactor cores, which can lead to improved safety and economy of these pivotal designs.
The simulations are running on P=4,096 up to P=32,768 processors of the IBM Blue Gene/P at the Argonne Leadership Computing Facility. The computations are based on the Nek5000 code, which simulates fluid flow, convective heat and species transport, and magnetohydrodynamics in general 2-D and 3-D domains. Nek5000 has been developed at Argonne National Laboratory under DOE’s Advanced Scientific Computing Research (ASCR) program. A singular feature is the code’s ability to scale to the large processor counts that characterize petascale computing platforms. Nek5000 was recognized with the 1999 Gordon Bell prize for algorithmic quality and sustained high performance on 4,096 processors of the ASCI-Red.
Results/Accomplishments
Researchers have simulated wire-wrapped fuel rods with 7- and 19-pin bundles. The current computations are some of the largest to date with Nek5000 and involve several hundred million gridpoints in unstructured domains. The scale of these computations has necessitated development of a new parallel strategy for solving the coarse-grid problem that is central to the efficiency of Nek5000’s multigrid solvers. The new solver employs algebraic multigrid, using Nek5000’s existing communication kernels, and results in sustained parallel efficiencies of ~60% for P=32,768 with only 3,700 points per processor.
The figure below shows a volume rendering of the axial velocity for a 19-pin bundle. To reduce computational costs, the simulations were performed using periodic boundary conditions in the axial flow direction, which allows turbulence to develop within a single wire-pitch. Figures of merit for these simulations include the channel-to-channel mixing and the pressure drop, both of which are significantly influenced by the presence of the side channels that are bounded by only two pins and a wall.
The second figure shows the velocity for a 7-pin bundle with three-wire pitches and a laminar inflow condition. This simulation illustrates that the flow becomes fully turbulent within a single wire pitch, which justifies the use of periodic boundary conditions when computing the turbulent flow field over lengths that span several wire pitches and many channel diameters. Calculations of this magnitude would not have been possible without resources at the BG/P scale.
University of Illinois and Argonne teams are working closely together to validate the core hydrodynamics large-eddy simulations by comparing highly detailed simulations in similar configurations. The university partners are performing simulations of coolant flow in a simplified geometry that allows them to resolve all turbulent motion with no modeling assumptions. These results are being compared to the Nek5000 computations, which simulate only the largest turbulent eddies in the flow. The validated Nek5000 results are being used to benchmark steady-state Navier-Stokes codes that employ turbulence models and to provide input to reactor design codes that require only coarse (mean flow) data. Visualization support for this project is being provided by the VisIt group at Lawrence Livermore National Laboratory (LLNL).
Future Efforts
Experiments indicate that low pin count results do not extrapolate to higher pin counts because of the edge channel effects. Succeeding simulations will involve more fuel pins, culminating in the design target of 217 pins. The group will be studying coolant flow in a variety of core subassembly designs in order to optimize reactor performance.
Related Video
Contact
Paul Fischer
Argonne National Laboratory
