Elsevier

Nuclear Engineering and Design

Volume 298, March 2016, Pages 218-228
Nuclear Engineering and Design

Benchmark exercise for fluid flow simulations in a liquid metal fast reactor fuel assembly

https://doi.org/10.1016/j.nucengdes.2015.11.002Get rights and content

Highlights

  • A EUROTAM-US INERI consortium has performed a benchmark exercise related to fast reactor assembly simulations.

  • LES calculations for a wire-wrapped rod bundle are compared with RANS calculations.

  • Results show good agreement for velocity and cross flows.

Abstract

As part of a U.S. Department of Energy International Nuclear Energy Research Initiative (I-NERI), Argonne National Laboratory (Argonne) is collaborating with the Dutch Nuclear Research and consultancy Group (NRG), the Belgian Nuclear Research Centre (SCK·CEN), and Ghent University (UGent) in Belgium to perform and compare a series of fuel-pin-bundle calculations representative of a fast reactor core. A wire-wrapped fuel bundle is a complex configuration for which little data is available for verification and validation of new simulation tools.

UGent and NRG performed their simulations with commercially available computational fluid dynamics (CFD) codes. The high-fidelity Argonne large-eddy simulations were performed with Nek5000, used for CFD in the Simulation-based High-efficiency Advanced Reactor Prototyping (SHARP) suite. SHARP is a versatile tool that is being developed to model the core of a wide variety of reactor types under various scenarios. It is intended both to serve as a surrogate for physical experiments and to provide insight into experimental results.

Comparison of the results obtained by the different participants with the reference Nek5000 results shows good agreement, especially for the cross-flow data. The comparison also helps highlight issues with current modeling approaches.

The results of the study will be valuable in the design and licensing process of MYRRHA, a flexible fast research reactor under design at SCK·CEN that features wire-wrapped fuel bundles cooled by lead-bismuth eutectic.

Introduction

Nuclear power plays an important role in power generation, producing about 16% of the total electricity worldwide. The rapidly growing energy demand suggests an even more important role for nuclear power in the future energy supply, as projected by the World Energy Outlook 2013 (IAE, 2013). Arguably, the accident at the Fukushima Daiichi nuclear power plant in Japan in March 2011 did have a minor effect on the future demand for nuclear power. Nevertheless, IAE (2013) calculates that nuclear power will be maintaining a 12% share of electricity generation globally by 2035, with expansion mainly in Asia. In Europe, in the Vision Report (SNE-TP, 2007) of the Sustainable Nuclear Energy Technology Platform (SNETP), a large role is attributed to the deployment of fast reactors. The preferred option is the sodium-cooled fast reactor, with the lead-cooled fast reactor as one of the two backups. Clearly, then, liquid metals will be important in the development of future nuclear energy technologies. A detailed overview of the status of fast reactor development is given in the IAEA (2012) report.

Thermal hydraulics is one of the key scientific factors in the design and safety analysis of liquid metal-cooled reactors. To solve thermal-hydraulic issues, nuclear engineers apply experiments, analytical and empirical correlations, and system thermal hydraulics codes or subchannel codes. Additionally, computational fluid dynamics (CFD) techniques are becoming increasingly integrated in the daily practice of the thermal-hydraulics researchers and designers. Roelofs et al. (2013b) summarize the current status and future challenges for CFD application to liquid-metal-cooled fast reactors. They show that for many liquid metal fast reactor thermal-hydraulic issues, the validation of CFD techniques is and will remain a key issue. In general, they underscore the simultaneous need for developments with respect to experiments including measurement techniques and numerical simulations.

Under the U.S. Department of Energy's International Nuclear Energy Research Initiative (I-NERI), Argonne National Laboratory (Argonne) collaborates with three Euratom members: the Dutch Nuclear Research and consultancy Group (NRG), the Belgian Nuclear Research Centre (SCK·CEN), and Ghent University (UGent) in Belgium on simulations of nuclear reactor core flows. The aim is to share data produced by the partners involved in order to systematically cross-verify fluid-dynamic simulations in liquid-metal-cooled nuclear reactor fuel assemblies. This collaboration focuses on code-to-code CFD comparisons in the absence of CFD-grade experimental data for wire-wrapped fuel assemblies.

Most liquid-metal-cooled fast reactor designs employ wire wraps as spacers between the individual pins in a rod bundle. Yet although many experiments have been performed, Roelofs et al. (2013a) clearly demonstrate that CFD-grade validation data is not available. New thermal-hydraulic experiments are under preparation in Germany, Italy, and the United States to fill this gap. To gain confidence in their employed Reynolds-averaged Navier–Stokes (RANS) approaches, the partners in this collaboration compare their results from RANS approaches with data from high-fidelity large eddy simulation (LES) performed at Argonne in a blind benchmark. Explanations of the various CFD modeling approaches are given by Roelofs et al. (2013a). These explanations basically show that direct numerical simulation and LES can provide high-fidelity reference data for comparison with more pragmatic RANS or hybrid approaches. The current paper describes the simulation efforts shared by the collaborating partners. Discrepancies or concerns with current prediction technologies are identified. Furthermore, an experimental plan for validation is under preparation taking into account the concerns that emerged from this collaboration.

The study is part of the code validation and verification approach developed in the licensing process of the Multi-purpose hYbrid Research Reactor for High-tech Applications (MYRRHA) under design at SCK·CEN (Abderrahim, 2012). MYRRHA is a flexible fast spectrum research reactor with wire-wrapped fuel bundles cooled by lead-bismuth eutectic. MYRRHA is identified as the European technology pilot plant for the lead-cooled fast reactor (LFR), which is one of the Generation IV reactor concepts (SNE-TP, 2010).

Argonne performed several wire-wrapped analyses as part of the Nuclear Engineering Advanced Modeling and Simulation (NEAMS) initiative (Pointer et al., 2008, Pointer et al., 2009, Smith et al., 2008). In particular, a simulation of a 7-pin wire-wrapped fuel bundle was performed with the LES code Nek5000. These calculations are used as the basis for a blind benchmark calculation among the collaborating partners. We emphasize that although pin counts in actual liquid metal assemblies are much higher, code comparisons in small assemblies are useful for building confidence in numerical and modeling practices at an early stage. Code-to-code comparisons in large assemblies are computationally expensive and present significant logistic challenges because of the large amount of data and the presence of multiple physical phenomena. Nevertheless, such comparisons in progressively larger assemblies are planned (see Section 6).

The present article describes the Nek5000 code (Section 2), briefly discusses the benchmark exercise (Section 3), and presents the comparisons of the benchmark datasets (Section 4) with the Argonne high-fidelity reference LES data. The results confirm previous findings (i.e., the good performance of the kω SST model in this type of modeling) and highlight the importance of consistent geometry in code-to-code comparisons.

Section snippets

Computational tools

Here we describe in detail SHARP and Nek5000 the tools used to perform the reference LES calculations.

Benchmarck exercise

The participants are collaborating on verification of numerical simulations pertaining to the flow in fuel pin bundles. These bundles typically incorporate spacers, such as wires or grids. Wire-wrapped pin bundles are notably used in liquid metal fast reactors. The presence of spacers complicates the flow considerably, and the lack of detailed experiments makes the validation of CFD simulations problematic.

As part of the NEAMS program, Argonne has performed several wire-wrapped analyses with

Comparison results

The data produced by the Nek5000 simulation was compared with five other calculations: three from NRG and two from UGent representing different computational and turbulence models:

  • 1.

    NRG-1: computed by using a kɛ cubic model

  • 2.

    NRG-2: computed by using the kω SST model

  • 3.

    NRG-3: computed by using a kɛ realizable model

  • 4.

    UGent-1: computed by using the kω SST model

  • 5.

    UGent-2: computed by using the kω SST model, with a more accurate surface representation

The NRG calculations were all performed with STAR-CCM+ (

Conclusions

The exercise described in this article has proven useful for discussing results of CFD codes across the entire spectrum of fidelity used by the SHARP toolkit and commercial codes such as STAR-CCM+. Comparison of the results obtained by the three institutions has shown good agreement at least for the cross-flow data (with an error below 5% for every calculation; UGent-2 and NRG-2 have maximum errors below 1.5%). Direct comparisons for the local velocity yielded mixed results, with the kω SST

Outlook

In addition to the work presented here, the participants in this benchmark performed a 19-pin wire-wrapped conjugate heat transfer benchmark. The geometry is related to the assembly design of MYRRHA. Results for this benchmark will be summarized in a forthcoming paper. Moreover, work on this topic in both the NEAMS and SESAME projects, as part of the verification activities of the computational fluid dynamic simulation of liquid metal assemblies.

Government license

The submitted manuscript has been created by UChicago Argonne, LLC, Operator of Argonne National Laboratory (“Argonne”). Argonne, a U.S. Department of Energy Office of Science laboratory, is operated under Contract No. DE-AC02-06CH11357. The U.S. Government retains for itself, and others acting on its behalf, a paid-up nonexclusive, irrevocable worldwide license in said article to reproduce, prepare derivative works, distribute copies to the public, and perform publicly and display publicly, by

Acknowledgments

The UGent contribution of the work described in this paper was funded by the Research Foundation – Flanders (FWO) with a Ph.D. fellowship and a postdoctoral fellowship.

The Dutch contribution of the work described in this paper was funded by the Dutch Ministry of Economic Affairs. Part of this work was supported by the FP7 EC Collaborative Project THINS no. 249337. This material was also based in part by work supported by the U.S. Department of Energy, Office of Science, under contract

References (23)

  • E. Merzari et al.

    Numerical simulation of the flow in wire-wrapped pin bundles: Effect of pin-wire contact modeling

    Nucl. Eng. Des.

    (2012)
  • F. Roelofs et al.

    Review of fuel assembly and pool thermal hydraulics for fast reactors

    Nucl. Eng. Des.

    (2013)
  • H.A. Abderrahim

    Multi-purpose hYbrid Research Reactor for High-Tech Applications a multipurpose fast spectrum research reactor

    Int. J. Energy Res.

    (2012)
  • Y.I. Chang et al.

    Advanced Burner Test Reactor Preconceptual Design Report, ANL-ABR-1 and ANL-AFCI-173

    (September 2006)
  • M.O. Deville et al.

    High-Order Methods for Incompressible Fluid Flow

    (2002)
  • P. Fischer et al.

    Petascale algorithms for reactor hydrodynamics

    J. Phys. Conf. Ser.

    (2008)
  • P. Fischer et al.

    Large eddy simulation of wire-wrapped fuel pins I: hydrodynamics in a periodic array

  • IAEA

    Status of fast reactor research and technology development

    IAEA Tecdoc 1691

    (2012)
  • IAE

    World Energy Outlook 2013

    (2013)
  • V.S. Mahadevan et al.

    High-resolution coupled physics solvers for analysing fine-scale nuclear reactor design problems

    Philos. Trans. A Math Phys. Eng. Sci.

    (2014)
  • E. Merzari et al.

    Benchmark exercise for fluid simulations in a liquid metal fast reactor core

    Argonne Report ANl/NE-19

    (2013)
  • Cited by (86)

    View all citing articles on Scopus
    View full text