Elsevier

Nuclear Engineering and Design

Volume 275, August 2014, Pages 312-321
Nuclear Engineering and Design

Benchmark analyses for EBR-II shutdown heat removal tests SHRT-17 and SHRT-45R

https://doi.org/10.1016/j.nucengdes.2014.05.027Get rights and content

Highlights

  • The IAEA EBR-II benchmarks SHRT-17 and SHRT-45R are analyzed with a 1D system code.

  • The calculated result of SHRT-17 corresponds well to the measured results.

  • For SHRT-45R ERANOS is used for various core parameters and reactivity coefficients.

  • SHRT-45R peak temperature is overestimated with the ERANOS feedback coefficients.

  • The peak temperature is well predicted when the feedback coefficient is reduced.

Abstract

Benchmark problems of several experiments in EBR-II, proposed by ANL and coordinated by the IAEA, are analyzed using the plant system code NETFLOW++ and the neutronics code ERANOS. The SHRT-17 test conducted as a loss-of-flow test is calculated using only the NETFLOW++ code because it is a purely thermal–hydraulic problem. The measured data were made available to the benchmark participants after the results of the blind benchmark calculations were submitted. Our work shows that major parameters of the plant are predicted with good accuracy. The SHRT-45R test, an unprotected loss of flow test is calculated using the NETFLOW++ code with the aid of delayed neutron data and reactivity coefficients calculated by the ERANOS code. These parameters are used in the NETFLOW++ code to perform a semi-coupled analysis of the neutronics – thermal–hydraulic problem. The measured data are compared with our calculated results. In our work, the peak temperature is underestimated, indicating that the reactivity feedback coefficients are too strong. When the reactivity feedback coefficient for thermal expansion is adjusted, good agreement is obtained in general for the calculated plant parameters, with a few exceptions.

Introduction

Benchmark problems concerning experiments in the EBR-II reactor, proposed by the Argonne National Laboratory (ANL) and coordinated by the IAEA, are analyzed by the one dimensional thermal–hydraulic system code NETFLOW++ and the neutronics analysis code ERANOS. The problems are the shutdown heat removal tests SHRT-17 and SHRT-45R conducted at the experimental fast breeder reactor EBR-II. Since the original part of the present work was conducted within the framework of a blind benchmark, no measured data were available until 17 January 2014. Measured data were made available to benchmark participants in February 2014. Test SHRT-17 is a natural circulation test after a loss-of-flow (LOF) event by tripping two pumps and scramming the reactor, and test SHRT-45R is a natural circulation test after an unprotected-loss-of-flow (ULOF) event. The calculated results for the former case are conducted under the blind benchmark, and the calculation for the latter case is re-evaluated result after the data was opened because we realized that there were several inaccuracies in our original results. A lot of data was measured in these tests. For the most important data, a comparison is shown in the present paper. In EBR-II temperatures were measured in two subassemblies by replacing the stainless wire spacers with thermocouples. These data show local temperature distribution in the instrumented subassemblies.

The NETFLOW++ code has been developed by Mochizuki (2010) and the code was verified using various data measured at mockups with water or sodium coolant, and validated using test data measured at the Monju and Joyo reactors (Mochizuki, 2007a). Since these reactors are so called loop-type reactors, every component needed for the analysis is already incorporated in the code. An inter-subassembly heat transfer model has been incorporated (Mochizuki, 2007b). As for a pool type FBR, the code was used for the IAEA natural convection benchmark analysis of PHENIX reactor (Mochizuki et al., 2013).

The EBR-II reactor is a very unique fast reactor which is classified as a hybrid-type reactor because of its heat transport system. The flow path from the main pump to the IHX through the reactor core is connected by piping, which is a similar configuration as in a conventional loop type reactor. However, the outlet of the IHX is not connected directly to the main pump but instead discharges into a large cold pool as shown in Fig. 1, and this arrangement is similar to the situation of pool type reactors. All components, the pump, the reactor core, piping for the primary heat transport system, and the IHX are in the sodium pool. Input for the NETFLOW++ code was created on the basis of the ANL benchmark documentation (Sumner and Wei, 2012). All reactor fuels, control rods, reflectors, blanket assemblies and others are modeled using 10 different kinds of fuel and reflector channels. The XX09 instrumented subassembly is separated from other channels.

The test result of SHRT-17 can be calculated using only the NETFLOW++ code because the reactor power transient is given in the benchmark problem. However, neutronic data and decay heat characteristics should be calculated in the case of the SHRT-45R. These are effective delayed neutron characteristics such as neutron life time, fractions of delayed neutrons and decay constants of the precursors. Feedback reactivity should be calculated too. For these purposes, the ERANOS v2.0 code is used.

The major objective of the analysis of the SHRT-17 is to adjust the thermal–hydraulic parameters for the calculation of SHRT-45R. If the SHRT-17 test cannot be calculated correctly, the calculation result SHRT-45R is likely to be unreliable. Although the benchmark is “blind”, some data, such as peak temperatures were reported in earlier publications by Dunn (1996) and Dunn et al. (2006). All measured data has been made available to the benchmark participants after the “blind” phase of the benchmark was over. The results in present paper in Sections 2 Calculation model and boundary conditions, 3 Calculation results correspond to the “blind” benchmark phase, and results in Section 4 correspond the phase after the measured data was made available.

Section snippets

Calculation model for thermal–hydraulics

The primary heat transport system of the EBR-II reactor is modeled on the basis of the ANL document (Sumner and Wei, 2012) as shown in Fig. 2. A simplified calculation model is shown in Fig. 3. The tank containing liquid sodium is modeled as a large diameter pipe with an equivalent inventory as the primary tank. Most pipes in the calculation model are divided into several nodes. The data of the heat transport system is made referring to above figures. Many kinds of assemblies are loaded in the

Steady state calculation

The steady state calculation was conducted with the time marching method for 10,000 s. During this period, all plant parameters converge to steady state values. As a result, the initial core inlet temperature converges to 625 K for the SHRT-17 test. This result is the same as the temperature measured in the test. In the case of SHRT-45R test, the initial reactor power of 60 MW was slightly higher compared to that of SHRT-17 at 57.3 MW power. Nevertheless, the initial core inlet temperature measured

Adjustment of negative reactivity

Since we understood the negative feedback and reference temperature were not properly chosen in the blind calculation, the mechanism of the net feedback was reconsidered after the measured data was made available. The feedback coefficients of the axial and radial core expansion were not changed. However, the feedback coefficient for coolant density is changed from −0.71 to −1.26 pcm/K and the feedback is calculated based on the coolant temperature on the core centerline rather than at the core

Conclusions

The following conclusions have been obtained through the blind benchmark calculations of the EBR-II reactor.

  • (1)

    The thermal hydraulic calculation model of the NETFLOW++ code is verified through the calculation of the SHRT-17 test. Good agreement is confirmed about the behavior of the peak sodium temperature in the instrumented channel.

  • (2)

    The temperatures at the various locations are predicted by the NETFLOW++ code without large discrepancies in general. Flow rates in the pump and the designated

Acknowledgements

The authors would like to express their gratitude to Mr. Stefano Monti of the Team Fast Reactor Technology Development at the IAEA for the specification and coordination of the EBR-II benchmarks. The authors would also like to acknowledge to the Argonne National Laboratory (USA) for making the data available for the benchmark exercise as well as assuring the technical coordination of the CRP.

Hiroyasu Mochizuki received a B.Sc. degree in 1972, followed by an M.Sc. and finally received a Ph.D. in 1979, all at the Tokyo Institute of Technology in Tokyo, Japan. His professional carrier started in 1978 at the Japan Atomic Energy Agency (JAEA) until 2006, as a scientist in the area of reactor thermal hydraulics. He was invited as a Visiting Professor at the Graduate School of Nuclear Power and Energy Safety Engineering at the University of Fukui from 2006 until 2008 and became full

Cited by (0)

Hiroyasu Mochizuki received a B.Sc. degree in 1972, followed by an M.Sc. and finally received a Ph.D. in 1979, all at the Tokyo Institute of Technology in Tokyo, Japan. His professional carrier started in 1978 at the Japan Atomic Energy Agency (JAEA) until 2006, as a scientist in the area of reactor thermal hydraulics. He was invited as a Visiting Professor at the Graduate School of Nuclear Power and Energy Safety Engineering at the University of Fukui from 2006 until 2008 and became full Professor in 2008 at the same university. From 2009 onwards, he is a Professor at the Research Institute of Nuclear Engineering, University of Fukui, in Tsuruga, Japan.

Kohmei Muranaka received a B.Sc. degree in Mechanical Engineering in 2014 at the University of Fukui, Japan, and started his education in M.Sc. in the graduate school of Nuclear and Energy Safety Engineering at the same venue.

Takayuki Asai received a B.Sc. degree in Mechanical Engineering in 2012 from University of Fukui, and started an M.Sc. degree in the graduate school of Nuclear and Energy Safety Engineering in 2013 at the same venue.

W.F.G. (Wilfred) van Rooijen received an M.Sc. degree from Delft University of Technology in the Netherlands in 2001, followed by a degree in Japanese Business, Language and Culture at Leiden University (The Netherlands) in 2002. In 2006 he received a PhD from Delft University of Technology in the area of nuclear reactor physics. He was an assistant professor at the Georgia Institute of Technology (USA) from 2007 to 2009, and since 2009 he is affiliated to the Research Institute of Nuclear Engineering at the University of Fukui in Japan, where he presently holds the position of Associate Professor in nuclear reactor physics.

View full text