Elsevier

Annals of Nuclear Energy

Volume 85, November 2015, Pages 93-114
Annals of Nuclear Energy

Design and performance of 2D and 3D-shuffled breed-and-burn cores

https://doi.org/10.1016/j.anucene.2015.04.007Get rights and content

Highlights

  • We report an in-depth study of breed-and-burn (B&B) reactor cores.

  • Principles for fuel, fuel assembly and core design of B&B systems were studied.

  • A concept of three-dimensional shuffling of fuel was introduced.

  • 3D-shuffling enables B&B operation with radiation damage below 350 DPA.

Abstract

The primary objective of this work is to find design approaches that will enable 3D fuel shuffling in stationary breed-and-burn (B&B) cores and to quantify the attainable reduction in peak DPA and change in additional performance characteristics going from conventional 2D to 3D fuel shuffling strategies. An additional objective is to establish the tradeoff between the minimum required DPA (displacements per atom) and average required burnup (fuel utilization) for B&B cores spanning a core power range from 1250 to 3500 MWth. It is found possible to design a B&B core fuelled with depleted uranium to have a peak radiation damage at or below 350 DPA when using 3D-shuffling. Relative to conventional 2D-shuffling, 3D-shuffling offers between 30% and 40% reduction in the peak DPA along with up to 30% increase in the average discharge burnup and, hence, in the depleted uranium utilization as well as significant increase in the core average and specific power density. Per DPA, the 3D shuffling option offers up to 60% higher uranium utilization. Even though 350 DPA is above the 200 DPA peak radiation damage HT9 steels were exposed to so far, it is below the 400 DPA advanced structural materials are expected to tolerate.

Introduction

Present-day light water reactors (LWR) extract only ∼0.6% of the potential fission energy of the natural uranium mined to make their fuel. About 90% of the unused uranium is depleted uranium (DU) and the remaining >9% is left over as used nuclear fuel (UNF). Based on the information from the 2006 US Department of Energy (DOE) UF6 cylinder information database, there is approximately 479,000 tons of depleted uranium stored in the form of UF6 in the US alone (Schneider et al., 2007). This material is currently located outside three large enrichment facilities (in Kentucky, Tennessee and Ohio) in metallic canisters. About 1330 tons of this material is 235U, with a concentration in the range of 0.2–0.3% 235U. The volume-averaged 235U-concentration of the US DU stockpile is 0.279%.

This material can, in principle, be utilized for energy production in fast reactors via, primarily, conversion of 238U by neutron capture to 239Pu. Direct fission of the 238U can contribute between 10% and 20% of the total fissions in a hard spectrum core. The conventional approach for attaining high uranium utilization is to use it to fuel fast breeder reactors (FBR) coupled with periodic fuel re-cycling. When the radiation damage accumulated by the fuel and its cladding threatens the mechanical integrity of the fuel rods, the fuel assemblies are discharged and reprocessed and the heavy metal is recycled back to the reactor core with the addition of depleted uranium makeup. Fuel reprocessing involves removal of the fuel cladding, removal of the fission products, addition of some DU to make up for the uranium that has been fissioned, fabrication of new fuel assemblies and reloading the fuel assemblies into the reactor core for another irradiation cycle. This process enables utilization (fission) of all the mined uranium minus inevitable losses in, primarily, the reprocessing operations.

Reprocessing of LWR fuel currently in use in both Europe and Japan uses the PUREX (Plutonium Uranium Redox EXtraction (Anderson and Asprey, 1947) process to create mixed oxide (MOX) fuel for use in thermal reactors. Recycling in any thermal spectrum reactor, however, cannot be done indefinitely since maintaining criticality requires increasing the Pu loading and this tends to make the coolant density reactivity coefficient positive. Moreover, as the conversion ratio of thermal reactors is smaller than 1, they need continuous feed of enriched uranium. Thus, France and Japan are contemplating a single recycle of plutonium through their thermal spectrum reactors. This will raise the level of uranium utilization up to only ∼1%.

Although reprocessing is currently in use in several countries for LWRs, there is significant objection in the United States and elsewhere to fuel reprocessing due to uncertainties regarding costs, and more importantly proliferation concerns. The proliferation concern is that the chemical or electrochemical processes that have been developed for fuel reprocessing could be used for the extraction and diversion of fissile materials for the clandestine production of nuclear weapons. Scientific arguments have been put forth that the reprocessing technology most suited for FBR applications (electrochemical) is non-proliferating by design (Hannum et al., 1997), meaning that the recycled materials could not be used to make a nuclear weapon without significant further processing (DeVolpi, 1986). These facts have so far had little effect on the positions of political opponents of fuel reprocessing.

Fast breeder reactors could, in principle, also operate without fuel recycling. That is, using a once-through fuel cycle as do the overwhelming majority of LWRs. Although the standard FBR discharge burnup of 10–15% FIMA (Fission of Initial Metal Atom) is two to three times higher than that of LWRs, the uranium utilization of a once-through FBR is not significantly different from that of a once-through LWR because the level of uranium enrichment required to achieve criticality in the FBR is often three times that required to fuel the LWR. A conventional FBR operating without reprocessing is thus not able to use fuel resources more efficiently or make any use of the untapped energy potential of the available DU stockpiles.

To enable a significantly higher utilization of uranium while using a once-through fuel cycle, a special class of fast nuclear reactors, collectively known as “breed-and-burn” (B&B) reactors, has been under consideration since the late 1950s. Theoretical studies have proven the principle of a traveling B&B deflagration wave through fertile material (so called “Traveling Wave Reactors”), but realistic designs utilizing this principle for power-producing cores have been difficult to achieve. The unique feature of a B&B reactor is that it can breed plutonium in low 235U containing uranium feed fuel and then fission a significant fraction of the bred plutonium without having to reprocess the fuel. In order to initiate the chain reaction, the initial fuel loading of the B&B core has to have an adequate amount of fissile material – either enriched uranium (EU) or plutonium with or without minor actinides. Thereafter, the B&B core is capable to continue its operation while only being fed with fertile fuel such as natural or depleted uranium. Eventually, the uranium utilization will approach the fraction of the loaded uranium that has been fissioned.

An extensive though most definitely not complete chronologic history of B&B research and development up until the start of the Terrapower efforts in 2006 is shown in Fig. 1. The first or most important publication from each identified B&B research group in Fig. 1 is given in references: (Feinberg and Kunegin, 1958, Fuchs and Hessel, 1961, Kouts et al., 1979, Atefi, 1980, Slesarev et al., 1983, Feoktistov, 1988, Goldin and Anistratov, 1992, Seifritz, 1995, Teller et al., 1996, Toshinsky, 1997, Akhiezer et al., 1999, Sekimoto and Ryu, 2000, Van Dam, 1998, Van Dam, 2000, Khizhnyak, 2001, Pilipenko et al., 2003, Fomin et al., 2005, Chen et al., 2005, Yarsky, 2005, Heidet and Greenspan, 2010, Gaveau et al., 2005, Gilleland, 2008). Since 2006, when Bill Gates announced support and funding for the idea (TerraPower, 2015), a large number of groups have started research on the topic. Some of the more recently published researched is presented in references (Pavlovich et al., 2007, Rusov, 2011, Hartanto and Kim, 2012, Osbourne et al., 2012, Choi et al., 2013, Si, 2013, Heidet and Greenspan, 2013a, Heidet and Greenspan, 2013b).

B&B reactors come in two basic flavors: Traveling Wave Reactors (TWR) and Standing (or Stationary) Wave Reactors (SWR) (Greenspan, 2012). TWRs are long cores of static fuel with a small, enriched, “starter” region typically on one axial end. The starter region initiates a breeding and burning wave that travels axially through the fertile fuel material toward the other end. To propagate the burning wave through low-enriched material such as DU, a very high level of fast neutron fluence is needed, which exposes core structures (like fuel rod cladding and fuel assembly duct wall steel) to excessive levels of radiation damage.

Current fast reactor structural steels – specifically the ferritic–martensitic HT9 steel, were exposed, so far, to a peak fast neutron fluence of 3.9 × 1023 n/cm2 (E > 0.1 MeV), corresponding to a radiation damage of ∼200 DPA (displacements per atom) (Toloczko et al., 1994). Realistic TWR designs using DU as fertile fuel require a minimum peak fast neutron fluence of at least 2.5 × 1024 n/cm2 (E > 0.1 MeV) (Tak et al., 2012), which corresponds to around ∼1200 DPA – six times above the currently accepted limit. In other estimates, the peak TWR discharge fluence is as high as 4.2 × 1024 n/cm2 (E > 0.1 MeV) (Kim and Taiwo, 2010). This situation makes TWRs impractical to implement until new materials or new design solutions for accommodating the excessive radiation damage can be found. One design solution investigated by Sekimoto and colleagues is to reclad the fuel rods with new steel every ∼10 years of operation (Nagata and Sekimoto, 2007, Nagata et al., 2009).

In SWRs, while burned fuel assemblies are discharged, remaining fuel assemblies are shuffled radially in the core and new fertile (DU) fuel assemblies are loaded. In this way, the breed-and-burn wave can stay stationary while the fuel “travels” through the wave. SWRs are more neutron-efficient than TWRs, because they lose a smaller fraction of neutrons to leakage or non-fuel absorption. Correspondingly, the minimum peak fast fluence required to sustain SWR B&B-type operation in an optimized core with DU feed fuel is ∼1 × 1024 n/cm2 (E > 0.1 MeV), corresponding to about 500 DPA (Heidet, 2010, Qvist, 2013). While this is significantly more likely realizable than the values for TWRs, it still requires the cladding steel to survive radiation damage level more than double the current experimental limit. Two approaches to tackle this problem have been pursued in previous work. Recently, the TerraPower design was adjusted to use natural rather than depleted uranium as feed fuel to lower the peak radiation damage down to 480 DPA (Gilleland, 2014). At UC Berkeley, researchers have looked into the possibility of B&B operation with periodic and limited fuel reconditioning (Greenspan and Heidet, 2011), as well the option to double-clad the fuel (Di Sanzo et al., 2011). With limited fuel reconditioning, fuel rods are taken out of the core once the cladding reaches 200 DPA, part of the fission products (primarily the gaseous ones) are removed and the cladding is replaced prior to fuel reuse in the reactor.

A different option, recently introduced and studied in detail in this paper, is to introduce three-dimensional fuel shuffling in B&B cores in order to minimize the peak DPA (Qvist and Greenspan, 2014). The axial peaking-factor (maximum value divided by the average value) for fluence, burnup and DPA in a typical SWR B&B is on the order of 1.7–2.7. If the axial DPA-profile could be flattened down to a peaking-factor close to 1.0, peak DPA levels could be brought down significantly.

The primary objective of this work is to find design approaches that will enable 3D shuffling in SWR cores and to quantify the attainable reduction in peak DPA and change in additional performance characteristics going from conventional two-dimensional to three-dimensional fuel shuffling strategies. An additional objective is to establish the tradeoff between the minimum required DPA and average required burnup (fuel utilization) for varying core sizes and power outputs.

Section 2 briefly reviews fundamentals of B&B reactor physics that will help understanding the special requirements and challenges faced in B&B core design. Sections 3 Fuel rod and assembly design for B&B cores, 4 Design considerations for 3D-shuffling of B&B cores covers the fuel rod and assembly design considerations of 2D and 3D-shuffled B&B cores respectively. Section 5 defines the fuel rod design that is used for all B&B cores in this study. Section 6 covers the design of the core and its structural components, particularly with regards to the 3D-shuffling options. Section 7 presents the approach of incorporating autonomous reactivity control (ARC) systems to tackle the challenges of inherent safety performance of large B&B cores. Sections 8 Core design of 2D and 3D-shuffling SWR B&Bs, 9 B&B fuel shuffling paths summarize the core design and shuffling scheme methodologies respectively. The results are summarized in Section 10, and Section 11 defines conclusions and the direction of future research.

Section snippets

The neutron balance concept and minimum required burnup

To estimate between what levels of average discharge burnup an SWR B&B core can operate at when at equilibrium, one can follow a batch of fertile material (metallic depleted uranium) from its introduction into the core until its discharge. When loaded, the fertile batch is a neutron sink, absorbing far more neutrons than it generates (k of a typical liquid–metal cooled core composition loaded with metallic DU fuel is ∼0.25). As neutrons are absorbed, the 239Pu concentration quickly builds up

Fuel rod design

The design of the fuel rod has a significant effect on the minimum required burnup and peak radiation damage via the following design variables – fuel material and density, fission gas plenum length and lattice pitch.

Conventional uranium dioxide fuel cannot sustain the B&B mode of operation in any configuration when fed by fertile fuel due to the poor neutron economy (Heidet and Greenspan, 2012, Qvist and Greenspan, 2015). A critical equilibrium cycle could theoretically be reached using either

Design considerations for 3D-shuffling of B&B cores

One of the objectives of this study is to design a B&B core in which 3D fuel shuffling can be implemented. The fuel assembly design for the 3D-shuffled B&B core presents a number of new challenges. The general approach taken is to axially subdivide the fuel assembly into several segments or “sub-assemblies”, which can be axially connected to form a fuel assembly equivalent to that used for the 2D-shuffled core design. When the reactor is shut down for shuffling, the sub-assemblies are

Suggested B&B core fuel rod design

Fig. 9, Fig. 10 show an axial and radial view respectively of the reference B&B fuel rod design using porous-plug fission gas venting and annular metallic fuel. The annular fuel is mechanically bonded to a 30 μm vanadium (or possible Zirconium) liner (Ryu et al., 2009), which acts as a diffusion barrier to avoid fuel/cladding chemical interaction (FCCI).

The ferritic/martensitic HT9 steel was chosen as the cladding material because it has the largest experimental database for high-dose neutron

Core and core structure design considerations

The B&B reactors considered in this study are assumed to be of the conventional large pool-type sodium fast reactors design with pumps and intermediate heat exchangers (IHX) located inside the reactor vessel. Due to the axial segmentation of fuel assemblies in the 3D-shuffling system, a number of modifications to the reactor vessel and its components have been implemented. A preliminary conceptual design of a 3D-shuffling core with 4 axial fuel assembly segments, not including components such

Safety systems and control

As mentioned in Section 2.1, the neutron economy required for B&B reactors operating with DU-feed fuel at equilibrium dictates the cores to be large with a low neutron leakage probability. This requirement presents a challenge to safety, since it is well known from previous studies that reactivity coefficients change for the worse with increasing core size and decreasing nominal neutron leakage probability (Wade and Fujita, 1987, Qvist and Greenspan, 2012) – the coolant temperature and void

General core design objectives and parameters

The primary objective of the B&B core design optimization enable equilibrium cycle operation on depleted uranium feed fuel while minimizing the peak DPA of the cladding steel, subjected to the design constraints as specified in Section 8.2. Optimal core design is searched for three power levels: 1250, 2500 and 3500 MWt (the core power levels considered covers the range of practical interest for B&B reactors); each for two fuel shuffling methods – 2D and 3D. The core volumetric power density for

Principles of B&B shuffling

Three-dimensional shuffling of 48 fuel batches (12 radial, 4 axial) can be arranged in a near-infinite number of ways. Complete optimization of the shuffling path is not amenable to brute-force numerical studies using the computationally intensive simulation tools presently in our use. Previous studies searched for optimal 2D SWR B&B shuffling using genetic algorithms that improve the shuffling path through thousands of iterations (Toshinsky et al., 1999, Toshinsky et al., 2000). Such methods

Results

The 1250 MWt cores were designed with 200 cm active fuel and an equivalent active core diameter of 300 cm, giving a core volume of 14.1 m3 and a volumetric power density of 88.4 MW/m3. The designs and calculation results for the 1250 MWt cores are given in Table 4, Table 5, respectively.

Because of the significantly lower radial power peaking factor of the 3D-shuffling core, the volume fraction (and mass) of fuel in the core is higher and the pitch-to-diameter ratio is lower. This gives the

Conclusions and future work

In this work, we defined and developed the engineering and physical principles of B&B core designs using both 2D and 3D shuffling of fuel. We have shown that it appears possible to design a fully optimized 3D-shuffling system for B&B operation using depleted uranium with a peak irradiation damage at or below 350 DPA. For conventional 2D-shuffled systems, it appears feasible to design systems that can operate on a feed fuel of depleted uranium and suffer a peak irradiation damage below 500 DPA.

References (86)

  • M.B. Toloczko et al.

    Irradiation creep and swelling of the US fusion heats of HT9 and 9Cr-1Mo to 208 dpa at ∼400°C

    J. Nucl. Mater.

    (1994)
  • V.G. Toshinsky et al.

    Multiobjective fuel management optimization for self-fuel-providing LMFBR using genetic algorithms

    Ann. Nucl. Energy

    (1999)
  • V.G. Toshinsky et al.

    A method to improve multiobjective genetic algorithm optimization of a self-fuel-providing LMFBR by niche induction among nondominated solutions

    Ann. Nucl. Energy

    (2000)
  • H. Van Dam

    Burnup waves

    Ann. Nucl. Energy

    (1998)
  • H. Van Dam

    Self-stabilizing criticality waves

    Ann. Nucl. Energy

    (2000)
  • A.I. Akhiezer et al.

    Propagation of a nuclear chain reaction in the diffusion approximation

    Phys. At. Nucl.

    (1999)
  • Anderson, H.H., Asprey, B.L., 1947. Solvent Extraction Process for Plutonium, US2924506, May...
  • ASME

    Boiler & Pressure Vessel Code, Section VIII, Division 1, Subsection A, Article UG-27

    (2005)
  • Atefi, B., 1980. Evaluation of the Breed/Burn Fast Reactor Concept (Ph.D. thesis). Massachusetts Institute of...
  • Barthold, W.P., 1974. Constrained Nuclear Design. In: Proc. of ANS meeting on Advanced Reactors; Physics, Design and...
  • Cahalan, J.E., Tentner, A.M., Morris, E.E., 1994. Advanced LMR safety analysis capabilities in the SASSYS-I and SAS4A...
  • Y.I. Chang

    Technical rationale for metal fuel in fast reactors

    Nucl. Eng. Technol.

    (2007)
  • Chen, X.N., Kiefhaber, E., Maschek, W., 2005. Neutronic model and its solitary wave solutions for a CANDLE reactor. In:...
  • H. Choi et al.

    A compact gas-cooled fast reactor with an ultra-long fuel cycle

    Sci. Technol. Nucl. Install.

    (2013)
  • A. DeVolpi

    Fissile materials and nuclear weapons proliferation

    Annu. Rev. Nucl. Part. Sci.

    (1986)
  • Di Sanzo, C., Cohen, J. Cisneros, T., Ludewight, B., Greenspan, E., 2011. An accelerator driven energy multiplier for...
  • M.T. Farmer et al.

    Generation IV Nuclear Energy System Initiative Pin Core Subassembly Design

    (2006)
  • Feinberg, S.M., Kunegin, E.P., 1958. Discussion comment, Ref. of Proc. Session B-10 ICPUAE 9(2), p....
  • L.P. Feoktistov

    An Analysis of a Concept of a Physically Safe Reactor

    (1988)
  • K. Fuchs et al.

    The possibilities for the operation of a natural uranium breeder reactor without fuel element preparation (in German)

    Kernenergie (East Germany)

    (1961)
  • General Atomics, 1974. Gas-cooled fast breeder reactor preliminary environmental report on 300-MW(e) demonstration...
  • Gilleland, J., 2008. Novel reactor designs to burn non-fissile fuels. In: Proc. Int. Congress Advances in Nuclear Power...
  • Gilleland, J., 2014. Global problems and global solutions. In: Presentation at the University of California Nuclear...
  • V. Goldin et al.

    Mathematical Modelling of Neutron-Nuclear Processes in Safe Reactor

    (1992)
  • E. Greenspan

    A phased development of breed-and-burn reactors for enhanced nuclear energy sustainability

    Sustainability

    (2012)
  • D. Hartanto et al.

    Physics and design studies for a small breed-and-burn reactor

    Trans. Am. Nucl. Soc.

    (2012)
  • Heidet, F., 2010. Maximum Fuel Utilization in Advanced Fast Reactors without Actinides Separation (Ph.D. thesis)....
  • Heidet, F., Greenspan, E., 2010. Breed-and-burn depleted uranium in fast reactors without actinides separation. In:...
  • F. Heidet et al.

    Neutron balance analysis for sustainability of breed-and-burn reactors

    Nucl. Sci. Eng.

    (2012)
  • F. Heidet et al.

    Superprism-sized breed-and-burn sodium-cooled core performance

    Nucl. Technol.

    (2013)
  • F. Heidet et al.

    Performance of large breed-and-burn core

    Nucl. Technol.

    (2013)
  • J.A. Horak et al.

    Irradiation Behavior of Restrained and Vented Uranium-2 w/o Zirconium Alloy

    (1962)
  • Idaho National Laboratory, 2013. Advanced Fuels Campaign 2013 Accomplishments, Idaho Falls, Idaho, US, Technical...
  • Cited by (20)

    • Feasibility of breed-and-burn reactor core design with nitride fuel and lead coolant

      2023, Annals of Nuclear Energy
      Citation Excerpt :

      Thus, very few reprocessing plants are currently available. To address these issues and to utilize the depleted and natural uranium effectively, scientists and nuclear engineers have studied the various concepts for fast breed-and-burn (B&B) reactors for many years, such as SWRs (Feinberg and Kunigen, 1958; Teller and Ishikawa, 1996; Teller et al., 1996; Rusov et al., 2015; van Dam, 2000; Shen et al., 2015; Lyman, 2021; Ellis et al., 2010; Gilleland et al., 2016; Hejzlar et al., 2013; Roberto and Juan-Luis, 2017), CANDLE by Sekimoto (Sekimoto and Ryu, 2000; Sekimoto et al., 2001; Sekimoto, 2001; Sekimoto, 2010; Sekimoto, 2005; Osato et al., 2018), and linear breed-and-burn fast reactors (Hartanto et al., 2016a,b, Kim et al., 2017) as well as concepts for reactors in B&B operating mode with minimum required burnup by Greenspan (Greenspan, 2012; Greenspan and Heidet, 2011; Heidet and Greenspan, 2012; Qvist and Greenspan, 2016; Zhang et al., 2017; Qvist et al., 2015). In SWRs, the nuclear fuel is burned and bred in various parts of the cores and then reshuffled to the high neutron flux region by remaining B&B wave stationary.

    • Design study of <sup>208</sup>Pb-Bi eutectic-cooled reactor with natural uranium as fuel cycle input with radial fuel shuffling

      2022, Annals of Nuclear Energy
      Citation Excerpt :

      The reactor has a low thermal efficiency of 32%-36% (Dragunov et al., 2015; Pioro, 2016). Furthermore, LWR only utilizes natural uranium as fuel less than 1% (Greenspan and Heidet, 2011a; Hoang et al., 2020; Kuwagaki et al., 2019; Qvist et al., 2015; Sekimoto and Yan, 2008; Zheng et al., 2016a). However, the Fast Breeder Reactor (FBR) with the open fuel cycle is not different from the LWR in utilizing natural uranium as fuel.

    • Development of methods for optimal assembly orificing design and application to a standing-wave Breed-and-Burn reactor

      2020, Nuclear Engineering and Design
      Citation Excerpt :

      Regardless, it may be extremely well-suited to B&B operation, and therefore is explored in this study. The reference core of this study is a prototypical large-core B&B similar to that designed by Qvist et al. (2015) with the help of the ADOPT code (Qvist and Greenspan, 2014). The core incorporates depleted U-Zr10 feed fuel clad in HT9, an aggressive assembly design, and a specific shuffling scheme which minimizes the peak radiation damage to the fuel cladding.

    • 3D in-core fuel management optimization for breed-and-burn reactors

      2016, Progress in Nuclear Energy
      Citation Excerpt :

      The thermal-hydraulics constraints accounted for in this study are: maximum core pressure drop: 0.9 MPa, (Qvist et al., 2015). maximum sodium velocity across the core: 12 m/s, (Fast Reactor Database 20, 2006).

    View all citing articles on Scopus
    View full text