19
Proceedings of ICAPP ‘08 Anaheim, CA USA, June 8–12, 2008 Paper 8026 Design Options for the Advanced High-Temperature Reactor Charles W. Forsberg Oak Ridge National Laboratory * P.O. Box 2008; Oak Ridge, TN 37831-6165 Email: [email protected] Tel: (865) 574-6783 P. F. Peterson University of California, Berkeley Berkeley, CA 94720-1730 Email: [email protected] Tel: (510) 643-7749 R. A. Kochendarfer Areva NP Lynchburg, VA 24506-0935 Email: [email protected] Tel: (434) 832-2990 File Name: ICAPP 2008 AHTR Design Options Paper Manuscript Date: January 15, 2008 Manuscript Number 8026 Session: Track 3: LMFR & Longer Term Reactor Programs 2008 International Congress on the Advances in Nuclear Power Plants (ICAPP’08) Anaheim, California June 8–12, 2008 The submitted manuscript has been authored by a contractor of the U.S. Government under contract DE-AC05-00OR22725. Accordingly, the U.S. Government retains a nonexclusive, royalty-free license to publish or reproduce the published form of this contribution, or allow others to do so, for U.S. Government purposes. _________________________ * Oak Ridge National Laboratory, managed by UT-Battelle, LLC, for the U.S. Department of Energy under contract DE-AC05-00OR22725. Charles Forsberg is retiring from ORNL and joining the Massachusetts Institute of Technology.

AHTR Options

  • Upload
    jammy36

  • View
    44

  • Download
    5

Embed Size (px)

Citation preview

Page 1: AHTR Options

Proceedings of ICAPP ‘08 Anaheim, CA USA, June 8–12, 2008

Paper 8026

Design Options for the Advanced High-Temperature Reactor

Charles W. Forsberg†

Oak Ridge National Laboratory* P.O. Box 2008; Oak Ridge, TN 37831-6165

Email: [email protected] Tel: (865) 574-6783

P. F. Peterson

University of California, Berkeley Berkeley, CA 94720-1730

Email: [email protected] Tel: (510) 643-7749

R. A. Kochendarfer

Areva NP Lynchburg, VA 24506-0935

Email: [email protected] Tel: (434) 832-2990

File Name: ICAPP 2008 AHTR Design Options Paper Manuscript Date: January 15, 2008

Manuscript Number 8026

Session: Track 3: LMFR & Longer Term Reactor Programs 2008 International Congress on the Advances in Nuclear Power Plants (ICAPP’08)

Anaheim, California June 8–12, 2008

The submitted manuscript has been authored by a contractor of the U.S. Government under contract DE-AC05-00OR22725. Accordingly, the U.S. Government retains a nonexclusive, royalty-free license to publish or reproduce the

published form of this contribution, or allow others to do so, for U.S. Government purposes. _________________________

*Oak Ridge National Laboratory, managed by UT-Battelle, LLC, for the U.S. Department of Energy under contract DE-AC05-00OR22725.

†Charles Forsberg is retiring from ORNL and joining the Massachusetts Institute of Technology.

Page 2: AHTR Options

Proceedings of ICAPP ‘08 Anaheim, CA USA, June 8–12, 2008

Paper 8026

†Charles Forsberg is retiring from ORNL and joining the Massachusetts Institute of Technology

Design Options for the Advanced High-Temperature Reactor C. W. Forsberg† P. F. Peterson R. A. Kochendarfer Oak Ridge National Laboratory U. of California, Berkeley Areva NP Oak Ridge, TN Berkeley, CA Lynchburg, VA Tel: (865) 574-6783 Tel: (510) 643-7749 Tel: (434) 832-2990 [email protected] [email protected] [email protected]

ABSTRACT–The Advanced High-Temperature Reactor (AHTR) is a new reactor concept under development that uses a solid high-temperature fuel and a liquid fluoride salt coolant. Depending upon the specific salt coolant that is selected, the liquid melting point of the coolant is between 350 and 500°C with boiling points typically at or above 1300°C. The very high volumetric heat capacity of the coolant and its other properties enable the development of compact modular reactors [100–1000 MW(t)] as well as high-power reactors [2000–4000 MW(t)] with passive safety systems. The base-case design with prismatic fuel blocks is described. The results of a series of trade studies that may significantly extend AHTR capabilities are also described. Studies of AHTRs with pebble-bed fuel have developed an AHTR concept with (1) significantly lower fuel cycle costs than traditional gas-cooled pebble-bed reactors and (2) uranium consumption per unit of electricity that is 64% of a light-water reactor. Studies of AHTRs with pin-type fuel assemblies for large power levels have developed core designs with the potential for significantly better actinide-burning capabilities than either light-water reactors or sodium-cooled fast reactors. A conceptual beyond-design-basis accident (BDBA) system has been developed that may allow failures of all decay-heat removal systems (except heat transfer to the ground) without significant release of radionuclides to the environment in a large [2000–4000 MW(t)] reactor. This is a significant enhancement in financial risk reduction compared to the present generation of reactors. The basis for these conclusions and the important technical uncertainties are described.

I. INTRODUCTION The goal of the Advanced High-Temperature Reactor (AHTR) program is the development of an economic high-temperature reactor with enhanced safety characteristics for the production of electricity and heat for industrial applications. These enhanced safety characteristics include passive safety systems and the capability to withstand a wide variety of beyond-design-basis accidents (BDBAs) in a large reactor system without significant radionuclide releases. The industrial applications1 include process heat for refineries, oil recovery, biomass conversion to liquid fuels, and hydrogen production. Depending upon the application, the peak reactor coolant temperature may vary from 700 to 950°C. With currently available code-qualified materials, an AHTR with a peak coolant temperature of ~700°C could be built. This near-term AHTR option would deliver heat at an average temperature of 650°C, the same average temperature as current modular high-temperature gas-cooled reactor (MHTGR) designs, and thus achieve the same power conversion

efficiency. With higher outlet temperatures the AHTR could deliver heat at significantly higher average temperatures than is feasible with MHTGRs. Fluoride salt coolants were originally developed and tested for molten salt reactors (MSRs). Two experimental MSRs were successfully built.2 In an MSR the fuel is dissolved in the coolant, whereas in the AHTR there is a solid fuel and clean coolant. The use of clean liquid salt coolants significantly simplifies reactor design because it eliminates many of the corrosion challenges associated with high concentrations of fission products in the coolant and avoids the challenges of fission products and actinides throughout the primary system. Like water, sodium, and helium coolants, the corrosion challenges in liquid salt systems are primary determined by the impurities in the coolant, with clean systems having low corrosion rates. This paper describes the baseline reactor design, including the system design, decay-heat removal system, and the BDBA system. The alternative

Page 3: AHTR Options

Proceedings of ICAPP ‘08 Anaheim, CA USA, June 8–12, 2008

Paper 8026

reactor core designs are also described, including their unique capabilities. The pebble-bed variant may enable the design of relatively small economic reactors and is the near-term option. The longer-term options include an AHTR with a pin-type fuel assembly and unique capabilities to burn actinides.

II. REACTOR DESCRIPTION The AHTR3–5 is a new reactor concept (Fig. 1) that combines, in a novel way, four established technologies: (1) coated-particle, graphite-matrix nuclear fuels (the same fuels used in high-temperature gas-cooled reactors (HTGRs); (2) Brayton power cycles; (3) passive safety systems and plant designs previously developed for liquid-metal-cooled fast reactors; and (4) low-pressure, transparent liquid-salt coolants. The reactor core physics are generally similar to those for HTGRs because the liquid salt coolant has a low neutron absorption cross section that can be comparable in magnitude to the salt’s contribution to neutron moderation. Reactor power is limited by a negative temperature coefficient, control rods, and other reserve shutdown systems. The option exists to include soluble neutron absorbers (rare earth fluorides) in containers with temperature-fusible openings (similar to fire sprinkler systems) above the reactor core in the coolant. This type of system would add neutron absorbers to the coolant should coolant overheating occur. Once activated, such systems could ensure shutdown independent of core temperatures (cold core after cooldown) or control rods. Several salts are being considered as potential reactor coolants. Table 1 shows the properties5–6 of some of these salts and other coolants. The baseline design uses a mixture of 7LiF and BeF2.4 The next three salts in Table 1 are potential alternatives. Isotopically separated lithium-7 is used in all the lithium salts because of its low nuclear cross section. The salts are mixtures of different fluorides to optimize their physical properties. The salts have melting points between 300 and 500°C and atmospheric boiling points typically above 1300°C. The reactor operates at near-atmospheric pressure, and at operating conditions, the liquid-salt heat-transfer properties are similar to those of water. The salt coolants are optically transparent6–7 with potentially some absorption in the far infrared. Heat is transferred from the reactor core by the primary liquid-salt coolant to an intermediate heat-transfer loop. The intermediate heat-transfer loop uses a secondary liquid-salt coolant8 to move the heat

to a thermochemical hydrogen production facility or to a turbine hall to produce electricity using a multi-reheat nitrogen or helium Brayton power cycle (with or without a bottoming steam cycle). Electrical efficiencies are expected to be near 50%. The AHTR capital costs have been estimated to be 50 to 60% those of a modular gas-cooled or liquid-metal-cooled reactor for equivalent electrical output.3 This is a consequence of economics of scale, higher potential efficiencies, and the higher volumetric heat capacity of liquid salts, which reduces equipment size relative to those of other reactor coolants (including sodium and water). Several different reactor core designs4 are being evaluated, including use of a prismatic-block fuel, a pebble-bed fuel, and a pin-type fuel. An engineering elevation view of the base-line prismatic-fuel design9 is shown in Fig. 2.

III. DECAY-HEAT REMOVAL SYSTEM Several alternative passive decay-heat removal systems have been investigated.4 The 2400-MW(t) baseline concept has a closed primary loop immersed in a tank containing a separate buffer salt that provides a heat sink for various transients. The temperature of the buffer-salt tank is near the lowest temperature in the primary system. If the intermediate heat-transport loop fails, the hot primary coolant exiting the primary heat exchanger transfers heat to the buffer salt through the uninsulated piping between the heat exchanger and the reactor core. As higher-temperature reactor coolant is recirculated back to the reactor, some of the higher-temperature coolant sweeps the inside of the reactor vessel and is cooled by the reactor vessel, which in turn is cooled by the buffer salt. If the primary pumps fail or are shut down, heat is also rejected from the primary system to the buffer salt through the Pool Reactor Auxiliary Cooling System (PRACS) heat exchangers. Each PRACS consists of a loop connecting the bottom reactor plenum to the top reactor plenum. The loop contains a fluidic diode and PRACS heat exchanger (PHX) to dump heat from the primary system to the pool. The fluidic diode is a no-moving-parts one-way valve that has high flow resistance in one direction and low flow resistance in the other direction. Typically the flow resistance in one direction is 50 or more times higher than that in the other direction. Such valves are used in a variety of chemical industries and have been used in some nuclear systems, such as the German sodium-cooled fast reactors.

Page 4: AHTR Options

Proceedings of ICAPP ‘08 Anaheim, CA USA, June 8–12, 2008

Paper 8026

ReactorPassive DecayHeat Removal

Hydrogen/Brayton ElectricityProductionReactor

Passive DecayHeat Removal

Hydrogen/Brayton ElectricityProduction

Fig. 1. AHTR layout for hydrogen production or electricity production.

TABLE I

Physical Properties of Coolantsa

Coolant Tmelt

(°C) Tboil

(°C) ρ

(kg/m3) Cp

(kJ/kg °C) ρCp

(kJ/m3 °C) k

(W/m°C) v ⋅106

(m2/s) Li2BeF4 (Flibe) 459 1430 1940 2.42 4670 1.0 2.9 59.5NaF-40.5ZrF4 500 1290 3140 1.17 3670 0.49 2.6 26LiF-37NaF-37ZrF4 436 2790 1.25 3500 0.53 31LiF-31NaF-38BeF2 315 1400 2000 2.04 4080 1.0 2.5 8NaF-92NaBF4 385 700 1750 1.51 2640 0.5 0.5 Sodium 97.8 883 820 1.27 1040 62 0.12 Lead 328 1750 10540 0.16 1700 16 0.13 Helium (7.5 MPa) 3.8 5.2 20 0.29 11.0 Water (7.5 MPa) 0 290 732 5.5 4040 0.56 0.13 aSalt compositions are shown in mole percent. Salt properties at 700ºC and 1 atm. Sodium-zirconium fluoride salt conductivity is estimated—not measured. The NaF-NaBF4 system must be pressurized above 700°C; however, the salt components do not decompose. Sodium properties are at 550ºC. Pressurized water data are shown at 290°C for comparison. Nomenclature used: ρ is density; Cp is specific heat; k is thermal conductivity; v is viscosity.

Page 5: AHTR Options

Proceedings of ICAPP ‘08 Anaheim, CA USA, June 8–12, 2008

Paper 8026

IHX modulesPump impeller

Control rods

Shielding plug

Reactor core

Radial reflector

Reactor vessel

Buffer salt tankCavity refractory insulationWater-cooled cavity liner

Refueling machine

Pump seal bowl

Reactor cover

Buffer salt free surface elev.

PHX w/ baffles

19.5 m

4.5 m8.0 m0 m 4 m

POWEROPERATION

Transfer gantry

REFUELING

Fuel transferchannel

Local fuelstorage hot cell

Neutron control assembly

DHX w/ baffles

Fig. 2. Cross section of a 2400-MW(t) Advanced High-Temperature Reactor with prismatic fuel. During normal operations the pumps push liquid salt through the core and attempt to push liquid salt through the PRACS loops; however, the flow is very limited because of the fluidic diodes. Upon loss of forced circulation (LOFC), decay heat is removed by natural circulation up through the reactor core and down through the PHXs and PRACS fluidic diodes so that it is dumped to the buffer salt. Heat removal from the buffer salt to the environment occurs primarily through Direct Reactor Auxiliary Cooling System (DRACS) heat exchangers. This is a closed natural-circulation heat-

transfer system consisting of two heat exchangers and associated piping. The DRACS heat exchanger (DHX) in the buffer-salt tank is at a lower elevation than the heat exchanger outside the reactor containment. Inside the natural-circulation decay-heat transport system, a fluid (e.g., a liquid salt, liquid metal, or other heat transfer fluid) absorbs decay heat from the DHX that is submerged in the pool salt. As the fluid is heated, its density decreases. The fluid flows by natural circulation through pipes to the upper heat exchanger outside the reactor containment. The heat is then rejected to ambient air, and the fluid is cooled. As the fluid is cooled, it

Page 6: AHTR Options

Proceedings of ICAPP ‘08 Anaheim, CA USA, June 8–12, 2008

Paper 8026

becomes denser and flows by natural circulation back through pipes to the DHX. This type of natural-circulation decay-heat transport system is used in some sodium-cooled fast reactors. Many variants of this system exist, including one that rejects the reactor decay heat to a water bath rather than to an air-cooled heat exchanger. With a two-salt system, the buffer salt does not have the nuclear requirement for low neutron absorption. This enables the selection of low-cost salts8 and the option of choosing a large pool with a high salt inventory that has multiple benefits because the buffer salt serves multiple functions, including the following. • Heat sink. The buffer salt provides a heat

sink during the first few hours after the loss of decay-heat removal, thus minimizing the size of the decay-heat system required in the long term to reject heat from the pool to the environment. The use of fluidic valves and the buffer-salt tank enable designs with small-temperature transients that minimize economic risks.

• Minimization of materials requirements.

The reactor vessel and components are bathed in a cooler salt with excellent heat-transport capabilities. This helps ensure primary system structural integrity during high-temperature transients. With appropriate design, the only components that operate at the peak temperatures of the reactor coolant are short lengths of pipe exiting the reactor, the pump heads, and the heat exchangers. This design minimizes the challenges associated with high-temperature materials.

• Radiation shielding. The buffer salt

provides shielding for the silo. • Operations and maintenance. The melting

points of the primary coolants (Table 1) are between 350 and 500°C. The buffer salt’s large heat capacity helps to prevent the primary salt and component temperatures from dropping below the freezing point of the salt. The low surface area of the buffer-salt tank simplifies system insulation. The salts are also transparent; thus, the components can be optically inspected7 from the buffer-salt tank during operations and plant shutdown.

• Accident protection and primary salt inventory control. The buffer salt contains neutron absorbers; thus, any failure of the primary system boundary allows this salt to enter the reactor and causes it to shut down. The buffer salt has a high solubility for most fission products and thus provides a secondary barrier against fission product releases. Last, in a BDBA with major structural failures (e.g., reactor vessel and buffer-salt tank failure), there is sufficient buffer salt to fill the silo above the level of the reactor core and thus ensure the reactor core is always covered with salt coolant. Under BDBA conditions, the buffer salt helps transfer heat from the reactor core to the silo wall.

The reactor system insulation is part of the silo with the silo cooling system to remove heat that is transmitted through the insulation. This system (below) also provides a secondary mechanism for removal of decay heat from the reactor system.

III. REACTOR CORE OPTIONS Three types of core designs are being considered for the AHTR,4, 10, 11 Each has specific advantages and disadvantages. The three fuel element designs [prismatic, pebble bed, and stringer (pin type)] are shown in Fig. 3.

III.A. Prismatic Fuel Blocks The initial development of the AHTR assumed that the fuel would be the graphite-matrix coated-particle prismatic-block fuel—a fuel originally developed for high-temperature gas-cooled reactors (HTGRs) in the United States. Two helium-cooled high-temperature reactors have been built using prismatic fuel blocks: (1) the Fort Saint Vrain Reactor (FSVR) that was built in Colorado and later decommissioned and (2) the operating Japanese High-Temperature Test Reactor (HTTR). In both cases, the reactor core is a three-dimensional stack of blocks. The FSVR prismatic fuel has been the baseline design of the AHTR. Changing from a gas coolant to a liquid coolant results in minor changes in the fuel, such as smaller coolant channels made possible by the better heat transport capabilities of liquid salts relative to those of helium.

Page 7: AHTR Options

Proceedings of ICAPP ‘08 Anaheim, CA USA, June 8–12, 2008

Paper 8026

Fig. 3. Alternative fuel configurations. The fuel consists of small microspheres coated with multiple layers that provide the cladding. The microspheres are incorporated into a graphite matrix that can be formed into different geometries. The result is a fuel that can operate under normal conditions with peak temperatures up to 1250°C. Under accident conditions, peak temperatures can approach ~1600°C for limited periods of time without significant fuel failure. No other demonstrated fuel has such high-temperature capabilities. Figure 4 shows the details of the prismatic fuel assembly used in the operating helium-cooled Japanese HTTR. In this fuel assembly, the microspheres are loaded into graphite annular fuel compacts that are then loaded into graphite sleeves (the fuel rods). The fuel rods are then loaded into holes in a graphite block. The fuel compact is an annular pellet designed to avoid the high temperatures in the middle of the fuel pellet and allow higher temperature operation. In terms of the AHTR, this fuel has five advantages: the graphite-based fuel is fully compatible with liquid salts, it is a high-temperature reactor fuel that matches the high-temperature capabilities of the coolant, the fuel has a strong negative Doppler coefficient for neutronic safety, the design of the fuel allows the core designer great

freedom in the fuel-to-carbon-to-coolant ratio, and it is fully developed. The neutronics of the prismatic-block AHTR with the lithium-beryllium liquid salt are almost identical3-4 to that of a HTGR. This is because of the very low volume fraction and absorption due to the salt. The baseline salt can also have a negative void coefficient. The alternative liquid-salt coolants have lower costs and several other potential advantages; however, these salts have slightly positive void coefficients and their parasitic neutron capture will reduce the achievable fuel discharge burnup. The preliminary analysis indicates that this is acceptable in terms of safety because the very large negative Doppler coefficient allows only a small rise in temperature if there is voiding. The addition of europium can further increase this negative Doppler coefficient. In this context, it is noted that many pressurized-water reactors (PWRs) have slightly positive void coefficients at low temperatures after refueling. This is not a safety concern and is acceptable to the U.S. Nuclear Regulatory Commission because the negative Doppler coefficient prevents temperature rises that could cause fuel damage.

06-032R2

Pebble BedStringerStringer

(not to scale)

Fuel Element

Prismatic

Fuel Block

Fuel

Ele

men

tsPl

ug U

nit

(~8

m)

(~14

m)

25 cm

1 m

Tie Bar(~1-cm diam)

Graphite Sleeve

Stainless Steel Pins

Page 8: AHTR Options

Proceedings of ICAPP ‘08 Anaheim, CA USA, June 8–12, 2008

Paper 8026

Fuel KernelPlug

Dowel Pin

GraphiteBlock

Annular CoolantChannel

Fuel Rod Fuel-HandlingHole

FuelCompact

GraphiteSleeve

High-Density PyC

0.92 mm

SiCLow-Density PyC

8 mm

39 mm

26 mm34 mm

Dowel Socket

Fuel Compact Fuel Rod

CoatedFuel

Particle

Fuel Block

360 mm

580 mm

ORNL DWG 2001-45R Fig. 4. Fuel design of the Japanese High-Temperature Test Reactor. While prismatic-block fuel assembles have major advantages, there are several disadvantages. Because of the high density of the liquid salt, graphite-matrix coated-particle fuel will float. Several methods to refuel prismatic AHTRs were evaluated10–11 that are viable; however, they add complexity. The prismatic fuel is relatively expensive to manufacture per unit of electricity that is generated. Simultaneously, investigations of alternative fuels led to several unexpected discoveries that make the other fuel forms potentially more attractive and indicate the possibility of a family of AHTRs with different core configurations for different applications.

III.B. Pebble-Bed AHTR The second type of ATHR fuel is pebble-bed fuel. A salt-cooled variant of this reactor is the pebble-bed AHTR (PB-AHTR). There is significant experience with this fuel type in gas-cooled reactors. Two helium-cooled pebble-bed reactors (PBRs) have been built, operated, and decommissioned in

Germany; a small pebble-bed test reactor has been recently built in China; and precommercial PBRs are being built in South Africa and China. The fuel consists of coated-particle fuel incorporated into pebbles. For helium coolant the typical diameter is ~6 cm. Pebble-bed fuels have lower fabrication costs than prismatic fuels because there is no complex geometry internal to the fuel. Figure 3 shows the German Thorium High-Temperature Reactor pebble-bed core. The vertical columns contain control rods. Pebble-bed reactors are refueled on-line. A slow continuous flow of pebbles circulates through the reactor core, with pebbles added at the top of the core and removed at the bottom. The pebbles go through the reactor core several times before being fully burnt. Extracted pebbles are then sent through a radiation detector to determine burnup as well as the disposition of the pebble as spent nuclear fuel (SNF) for disposal or for recycle back to the core for additional burnup.

Page 9: AHTR Options

Proceedings of ICAPP ‘08 Anaheim, CA USA, June 8–12, 2008

Paper 8026

Studies show that an equivalent PB-AHTR is viable.12–14 However, there is one major difference relative to the gas-cooled variant—the pebbles will float. Scaled experiments at the University of California–Berkeley (UCB) using water with polyethylene spheres that match the pebble-to-liquid density ratio and Reynolds and Froude numbers indicate that a pebble-bed reactor with floating pebbles is viable and can be refuelled.13 The scaled experiments confirmed smooth flow of pebbles through a simulated reactor core, refueling chutes, and other restrictions. UCB design studies, described in greater detail in a companion paper,14 have developed a modified PB-AHTR with potentially superior performance based on the high-performance cooling capabilities of liquid-salt coolants relative to helium. The 900-MW(t) PB-AHTR (Fig. 5) has one central hexagonal, removable pebble channel assembly (PCA) surrounded by six PCAs in a ring. Each of the seven PCAs has 19 small pebble channels (19.8 cm diam.). Each channel is a separate pebble bed. The pebble diameter is 3 cm. In the outer ring of PCAs, one of the 19 small pebble channels in each PCA is used for the control rods. Figure 6 shows the vertical cross section. This design has several major advantages. • Graphite separation. Reactor physics

requirements for negative void reactivity determine the fuel-to-carbon ratio in pebble-bed reactors. In a traditional gas-cooled PBR, the pebbles are primarily graphite with a very low uranium content. The low uranium content of the pebbles results in large volumes of fuel per unit of electricity produced. In the UCB design, over half of the graphite is in the PCA structure, not the pebbles. The pebbles have a much higher uranium loading and power density than in traditional gas-cooled PBRs; consequently, the number of fuel pebbles that must be manufactured and the volumes of SNF to manage are reduced substantially. This alternative design is viable because the use of a liquid salt coolant with its superior heat transfer capabilities allows for much higher power densities.

• Multi-channel fueling. A major design

constraint in gas-cooled PBRs is the random characteristics of the pebble bed that do not allow variable enrichment across the reactor core. This results in power peaking in the center of the core and a small fraction of the

fuel operating close to its temperature limits. With the multi-PCA pebble-bed design, the highest-burnup pebbles are recycled to the center PCA with the highest neutron flux while the lower-burnup pebbles are recycled through the six surrounding PCAs. The recycle of higher burnup pebbles to the central channel flattens the neutron flux and results in a more even power density with lower peak fuel temperatures.

• Graphite conservation. Both fuel and

graphite have limited lifetimes in a reactor core. However, the graphite lifetime is significantly longer than the fuel lifetime. In a gas-cooled PBR, the graphite in the pebbles increases waste volumes. In the multi-channel PB-AHTR, over half of the graphite is in the core (not the fuel) and is infrequently replaced. In addition, the disposal of irradiated graphite is significantly less difficult and expensive than the disposal of SNF.

These design changes have highly favorable economic implications for the fuel cycle. The higher fuel loading per pebble reduces the number of pebbles that must be fabricated and lowers fuel manufacturing costs. The higher energy output of each pebble reduces SNF storage, transportation, and disposal. The design also enables in-core fuel management. Historically gas-cooled PBRs had the advantage of uniform burnup of all SNF that improved fuel economy. This is in contrast to light-water reactors (LWRs) where the fuel burnup varies from the middle to the end of the fuel assembly. However, gas-cooled PBRs had no way to spatially vary fuel enrichment and burnup across the reactor core—it was uniform. In contrast, the fuel enrichments and burnups vary across an LWR reactor core—a fuel management characteristic that improves LWR fuel-cycle economics. With the multi-channel PB-AHTR and high-burnup pebbles in the center, the PB-AHTR core now has variable fuel properties that can be used to improve fuel cycle economics.

Page 10: AHTR Options

Proceedings of ICAPP ‘08 Anaheim, CA USA, June 8–12, 2008

Paper 8026

Fig. 5. Reactor core cross section of 900-MW(t) PB-AHTR. A comparison between PB-AHTR and PWR fuel cycles was undertaken. The PWR uses 4.5% enriched fuel, has a burnup of 50,000-MWd/ton, and has a power conversion efficiency of 33%. The PB-AHTR uses 10% enriched fuel, has a fuel burnup of 129,000 MWd/ton, and has a power conversion efficiency of 46%. The PB-AHTR uranium consumption and uranium enrichment requirements are, respectively, 64.2% and 86.2% those of a PWR per unit of electricity. Other neutronics studies have confirmed that the PB-AHTR can achieve negative void reactivity when operated with deep burn fuels composed of pure plutonium or pure transuranics. These studies used the lithium-beryllium fluoride salt as the coolant. No detailed capital-cost comparisons have been undertaken. However, size comparisons of a gas-cooled PBR and a PB-AHTR have been undertaken. This is shown in Fig. 7, which compares a

900-MW(t) PB-AHTR with a 400-MW(t) gas-cooled PBR. Table II provides comparisons with other reactors. There are major technical challenges and uncertainties. The PB-AHTR coated-particle fuel operates at peak fuel temperatures several hundred degrees lower than the peak fuel temperatures in gas-cooled PBRs. This is a consequence of the better heat transfer performance of liquids compared with gases. However, the power densities are significantly higher and there is limited experience in operating coated particle fuels under these conditions. Only limited design studies have been conducted. While the MSR programs partly developed prototype equipment and the PB-AHTR operates at the same temperatures as the Molten Salt Reactor Experiment (MSRE), an 8-MW(t) test reactor, there was limited experience. These issues, and the development path for the PB-AHTR, are discussed further in a companion paper.14

Page 11: AHTR Options

Proceedings of ICAPP ‘08 Anaheim, CA USA, June 8–12, 2008

Paper 8026

Fig. 6. Vertical cross section of a 900-MW(t) PB-AHTR.

Page 12: AHTR Options

Proceedings of ICAPP ‘08 Anaheim, CA USA, June 8–12, 2008

Paper 8026

TABLE II

Reactor System Comparisons with the General Electric Economic Simplified Boiling Water Reactor and the Gas Turbine-Modular Helium Reactor

Reactor Parameter ESBWR GT-MHR PB-AHTR

Power, MW(t) 4,700 600 900

Power, MW(e) 1550 286 411

Vessel height, m 27.7 25 11

Vessel diameter, m 7.1 8.4 6

Vessel pressure, MPa 7.00 7.5 0.80

Vessel temperature, °C 286 495 650

Vessel power density, MW(e)/m3 1.41 0.21 1.32

Vessel mass, metric tons 1050 840 150

Reactor head mass, metric tons 150 490 50

Reactor building height, m 70 81 35

Core power density, MW(t)/m3 4.8 ~25

Peak coolant temperature, °C 850 700

Peak fuel temperature during accidents, °C 1600 1200

III.C. Stringer-Fuel Core The third type of HTGR fuel is the stringer fuel. Like with other high-temperature fuel types, there is a parallel salt-cooled reactor version. Most of the experience for stringer-type fuel is in the United Kingdom where 14 Advanced Gas-Cooled Reactors (AGRs) were built and are currently operating. The AGRs are graphite-moderated, carbon dioxide–cooled reactors that use stainless steel–clad uranium dioxide fuel assemblies. The graphite AGR core has vertical holes for fuel stringers (Fig. 3). A stringer consists of multiple fuel assemblies, neutron moderator sections, radiation shielding, pressure seals, and other components. The 1-m-long fuel assemblies have a graphite sleeve, 36 stainless steel−clad fuel pins with uranium dioxide fuel pellets,

and a 1-cm nimonic-alloy-PE16 tie bar that goes through each fuel assembly and holds them together as a single unit on a stringer. The graphite sleeve provides a gas flow channel, serves as part of the assembly with the grid structure that holds the fuel pins in the proper geometry, and provides some radiation shielding to reduce the rate of radiation damage to the permanent graphite in the reactor core. The sleeve is not part of the SNF and is separated from the SNF pins for the purposes of disposal. It is not reused. The British AGRs operate at high pressures and refuel online. Consequently, above the fuel assemblies is a plug unit that goes through the AGR reflector block, insulation system, and shielding. It connects to the pressure boundary of the reactor.

Page 13: AHTR Options

Proceedings of ICAPP ‘08 Anaheim, CA USA, June 8–12, 2008

Paper 8026

Fig. 7. Scaled drawing of 900-MW(t) PB-AHTR and 400-MW(t) gas-cooled PBR.

Page 14: AHTR Options

Proceedings of ICAPP ‘08 Anaheim, CA USA, June 8–12, 2008

Paper 8026

The AGR core design and stringer-type fuel4, 10 is potentially applicable to liquid-cooled AHTRs with appropriate modifications. The major modification would be to replace the stainless steel clad of the AGR with a higher-temperature clad that is compatible with liquid salts. There are multiple high-temperature options including silicon carbide and several high-temperature metallic alloys such as molybdenum alloys. There are also other options such as oxide-dispersion-system (ODS) steels if the peak temperatures are lowered through the use of lower melting point salts. If silicon carbide were the replacement clad, the short fuel bundles would remain because of the difficulties in fabricating long silicon carbide rods. If a metallic alloy were the replacement clad, the fuel bundle might be the full length of the reactor core. The other modification is that the plug unit above the fuel can be eliminated because the AHTR is a low-pressure reactor. The British initiated a development program for an advanced AGR that used carbon dioxide cooling but operated at much higher temperatures. Most of the effort was to develop coated-particle fuel with a clad of silicon carbide. A smaller effort was undertaken to develop the silicon carbide clad with uranium dioxide fuel pellets. This would create a fuel assembly that could withstand very high temperatures under accident conditions. The clad, not the uranium dioxide, limits fuel temperatures in similar LWR fuels. The development program included extensive testing of the clad. While the initial results were highly promising, the program was cancelled when it was decided to adopt helium cooling for future reactors—a coolant that eliminated the need for a clad to protect the fuel against chemical corrosion. More recently, there has been a renewed interest in silicon carbide–clad fuels for (1) helium-cooled fast reactors in France and Japan and (2) LWRs in the United States. From an economic perspective, there are strong incentives to use uranium dioxide fuel pellets in a high-temperature clad because of the potential for very low fuel fabrication costs. The available evidence11 is that a suitable cladding could be developed if there were sufficient incentives. The core configuration for such a fuel assembly is shown in Fig. 8, where each fuel assembly is surrounded by graphite. The geometry and power density are somewhat similar to a CANDU reactor. The CANDU reactor has bundles in tubes in a tank filled with heavy water that is a neutron moderator. The stringer AHTR has fuel bundles in a matrix of graphite that is a neutron moderator. In both reactors, there is a thermal neutron flux in the moderator. The neutron flux becomes increasingly harder from the

outside to the inside of the fuel bundle because the outer rows of pins selectively absorb thermal neutrons. However, there are major differences in the reactor physics. Heavy water cools the CANDU fuel assemblies and is a good moderator; thus, the neutron flux is only somewhat harder in the center of the fuel bundle than on the outside. Liquid salts are poor moderators; thus, the neutron flux becomes a soft fast spectrum in the center of the fuel assembly. Initial AHTR stringer-fuel neutronic studies4 replaced some of the fuel pins in the assembly with graphite rods to help create a thermal neutron flux in the center of the fuel assembly. The result was a reactor core with behavior somewhat similar to the prismatic-block AHTR with the advantages of a pin-type fuel assembly and the AGR engineering experience of refueling pin-type fuel assemblies with a graphite matrix in the reactor core. The low pressure of an AHTR would significantly simplify the engineering. It was then recognized at Areva15 that this reactor (1) had a unique characteristic because of the unique combination of fuel, moderator, and coolant—fast spectrum islands within a thermal neutron reactor—and (2) this unique characteristic could be used as a new approach to actinide burning. The reactor was then assessed for its capability to burn actinides. The Areva evaluations primarily used the sodium-zirconium fluoride salt as a coolant with some evaluations using the lithium-sodium-beryllium fluoride salts as coolants. The latter salt has a significantly lower melting point. Because each row of pins has a somewhat different neutron spectrum, there is the choice to place specific actinides (plutonium, neptunium, etc.) in those rows where the neutron flux characteristic was best suited to burning the particular actinide. The initial studies by Areva indicated that the reactor had potentially superior abilities for actinide burning compared with traditional fast reactors. The potential advantages were partly a result of the neutron spectrum and partly because of the fact that higher amounts of actinides can be loaded into the AHTR than a sodium-cooled fast reactor without encountering unacceptable reactivity coefficients. Table III shows the results of the initial assessment15 of the stringer-fuel AHTR for actinide burning relative to several other reactors. Actinide burning is measured in terms of actinide destruction per unit of electricity that is produced. The results show the potential for more rapid burning of actinides per unit of electrical output relative to PWRs, sodium-cooled fast reactors, and other high-temperature reactors.

Page 15: AHTR Options

Proceedings of ICAPP ‘08 Anaheim, CA USA, June 8–12, 2008

Paper 8026

Fig. 8. Unit cell of AHTR with stringer fuel assembly with five rows of pins in graphite matrix.

TABLE III

Actinide Reduction Capability of Various Reactors (Kilograms Destroyed per Terra-Watt Hour of Electricity)

Reactor type 239Pu 240Pu 241Pu 237Np 241Am

Sodium fast reactor 50 5 20 10 <10

Deep burn MHTGR 55 30 5 5

Stringer-fuel AHTR 75 20 50 40

PWR 50

Page 16: AHTR Options

Proceedings of ICAPP ‘08 Anaheim, CA USA, June 8–12, 2008

Paper 8026

The most important isotopes in terms of actinide burning are 239Pu, 241Pu, 241Am, and 237NP. There are three potential goals for actinide burning. The first goal is the destruction of weapons-usable isotopes, thus the incentives for rapid destruction of fissile plutonium isotopes. A second goal is to improve repository performance. In terms of repository performance, as measured by reduction of long-term releases of radionuclides from the repository, neptunium is the critical isotope because of its long half-life and potential to migrate more rapidly in groundwater than other actinides. Consequently, for improved repository performance, there are incentives to destroy 237Np and the 241Pu that decays to neptunium. The last goal is to reduce the heat load in the repository from radionuclide decay and thus increase the amount of waste that can be disposed of in a unit area of the repository. The primary actinide of concern in terms of heat generation is 241Am and the 241Pu that produces it. Only limited work has been done on the use of stringer-type fuel assemblies for actinide burning. The conclusions are highly favorable, but additional work is required. Given the massive British experience with the AGR and the applicability of much of that engineering experience to a stringer-fuel AHTR, the primary technical challenge for such an AHTR is likely to be the development and qualification of the high-temperature clad.

IV. BEYOND DESIGN BASIS ACCIDENTS The unique characteristics of liquid-salt coolants may enable large AHTRs up to 4000 MW(t) to withstand BDBAs. The work is at an early stage of development16 and thus there are major uncertainties. The general design strategy is described herein. In a BDBA, it is herein assumed that the PRACS and the silo cooling systems have failed and that significant structural failures (vessel failure, etc.) have occurred. The reactor is in an underground silo, and effectively all above-grade components can be destroyed. The ultimate heat sink is the ground that then conducts heat to the atmosphere. There are major incentives for development of a reactor that cannot catastrophically fail with the potential release of large quantities of radionuclides to the environment. One of the reasons for development of modular high-temperature gas-cooled reactors (MHTGRs) is that these reactors have this characteristic. However, the cost to obtain this benefit has limited the size of the reactor. For the prismatic-core MHTGR, the maximum size reactor that can withstand this accident without major fuel

failure is ~600–800 MW(t). In an MHTGR BDBA, decay heat must be moved from the center of the reactor core to the vessel boundary by conduction and radiation, from the reactor vessel to the silo, and from the silo through the soil to the atmosphere. The driving force is the temperature difference between the hottest fuel and the environment. The failure temperature of the fuel (~1650°C) determines the peak temperature. This process involves a large temperature drop, transferring heat from the graphite-matrix coated-particle fuel to the graphite reflector and to a thick-wall pressure vessel. In a BDBA, MHTGR fuel may also be exposed to chemical attack from air or steam. In an AHTR BDBA without vessel failure, natural circulation of the salt within the core creates nearly uniform temperature within the reactor vessel and buffer-salt tank (Fig. 9). The large temperature drop seen in the MHTGR from the center of the reactor core to the reactor vessel is eliminated. Equally important, the large heat capacity in the system can absorb the initial decay heat uniformly until the rate of decay heat generation has dropped significantly. Because the fuel remains submerged under the surface of a pool of chemically inert salt, chemical attack by air and water is also precluded. The liquid coolant also implies that the AHTR has much higher peak reactor vessel and buffer-salt tank temperatures after an accident and that most of this system is at this much higher temperature. The higher relatively-uniform buffer-salt tank temperature provides a larger temperature drop to move heat from the buffer-salt tank to the silo and onto the environment. This greatly increases the ability to remove decay heat, which implies more power output from the reactor core while avoiding fuel failure in a BDBA. If there is buffer-salt tank failure, salt exits the tank. However, there is sufficient buffer salt to fill the space in the silo between the buffer-salt tank and the silo while keeping the reactor core covered. The silo becomes a leak-proof container because the salt freezing temperature (depending upon the buffer salt) is between 300 and 500°C. If the salt finds a crack in the silo, it progresses through the crack until it freezes. In effect, the high-temperature salt provides a self-sealing mechanism that is not available for lower-melting-point coolants such as water. Heat from the silo conducts through adjacent structures to the soil and the environment.

Page 17: AHTR Options

Proceedings of ICAPP ‘08 Anaheim, CA USA, June 8–12, 2008

Paper 8026

03-149R4

MHTGRHelium

600 MW(t)Conduction

AHTRLiquid Salt

>2000 MW(t)Tank Temperature

AHTRLiquid Salt

>2000 MW(t)Silo

ReactorCoolantPower

Limitation

Passively Cooled Wall

(All Reactors)

Uniform Tank

Temperature (Liquid)

Higher Tank Temperature

(Low Pressure)

Thick Vessel Wall

Thin TankWall

FrozenSalt

ReactorVessel

Fig. 9. Accident evolution for MHTGR, AHTR with intact buffer-salt tank, and AHTR with failed buffer-salt tank. The silo contains two BDBA systems (Fig. 10). The silo wall consists of low-cost, thick steel rings that are similar to those used in the mining industry to line deep mine shafts and prevent their collapse. The rings conduct heat up the silo wall and distribute it above the coolant salt layer. The second component is an annular ring of a low-cost, solidified BDBA salt. As the temperature of this salt increases, the BDBA salt melts, flows into the silo, and floods the silo to a higher level. The melting, heating, and boiling of the BDBA salt (1) provides a significant source of thermal inertia and (2) a heat transfer liquid to more efficiently transfer heat from the buffer-salt vessel to the entire vertical height of the silo. Under severe conditions there may be localized fuel failure because of localized flow blockages. As discussed earlier, the salt boils at lower temperatures than the failure temperature of the fuel; thus, localized boiling of the salt can help provide local cooling. If there is localized fuel failure, the liquid-salt coolant above the reactor core provides a secondary independent barrier to radionuclide escape. The chemistry of liquid-salt coolants is radically different from helium, sodium, or water coolants. These salts were originally developed for MSRs,2 which were originally developed for the Aircraft Nuclear Propulsion Program with the goal of providing very

high-temperature heat to operate aircraft jet engines. One of the MSR salt requirements was that the actinides and most of the fission products dissolve and remain in the salt at very high temperatures. Most fission products (including cesium and iodine) and all actinides escaping the solid AHTR fuel are highly soluble in the salt and will remain in the salt at very high temperatures, while the noble metal fission products are reduced to metal and precipitate on primary loop surfaces. Only the noble gases and tritium can easily escape. In effect, the liquid salts are a chemical containment system for actinides and fission products that is separate and independent of any mechanical containment system. The time evolution of an AHTR BDBA starts with a jumbled core with circulating liquid salt transferring heat outward toward the silo. The heat wave progresses outward. When decay heat becomes less than the rate of heat loss, the system begins to cool. Ultimately, the furthest out salt will freeze. As time progresses and the reactor cools down, the fuel is ultimately frozen in salt. The potential to transfer heat from the reactor core to the ground efficiently and seal any leaks in the silo is a direct consequence of the salts used in the system. There are significant uncertainties; however, the unique characteristics of a salt that freezes near 350°C to ensure silo sealing and a salt that dissolves fission products creates new options in BDBA systems.

Page 18: AHTR Options

Proceedings of ICAPP ‘08 Anaheim, CA USA, June 8–12, 2008

Paper 8026

Fig. 10. Beyond-design-basis accident.

V. CONCLUSIONS The AHTR is a new type of reactor with its own unique advantages and disadvantages. A series of trade studies have refined the design and discovered new characteristics of the AHTR that were previously unrecognized. These characteristics indicate the potential for the AHTR to be an economic high-temperature reactor. The studies also identified three potentially unique characteristics of the AHTR. The first discovery is the potential for a PB-AHTR with very good fuel economy in terms of uranium resource utilization with the capability to operate with deep-burn transuranic fuels, while minimizing the fuel fabrication costs and SNF management costs. The second discovery is the potential for superior capabilities in actinide burning. The last discovery is the possibility of designing a very large AHTR with the BDBA capability to withstand massive damage, including destruction of most of the station. All of these results are consequences of the choice of reactor coolant. Significant work will be required to confirm these initial results.

REFERENCES 1. C. W. FORSBERG, “Nuclear Energy for a Low-

Carbon-Dioxide-Emission Transportation System with Liquid Fuels,” Nuclear Technology (in press).

2. Nucl. Appl. Technol. (entire issue on molten salt

reactors), 8(2) (1970). 3. C. W. FORSBERG, “Goals, Requirements, and

Design Implications for the Advanced High-Temperature Reactor,” ICONE14-89305, CD-ROM, 14th International Conference on Nuclear Energy, American Society of Mechanical Engineers, Miami, Florida, July 17–20, 2006, American Society of Mechanical Engineers (2006).

4. D. T. INGERSOLL, C. W. FORSBERG, and

P. E. MACDONALD, Trade Studies on the Liquid-Salt-Cooled Very High-Temperature Reactor: Fiscal Year 2006 Progress Report, ORNL/TM-2006/140, Oak Ridge National Laboratory, Oak Ridge, Tennessee (2006).

Page 19: AHTR Options

Proceedings of ICAPP ‘08 Anaheim, CA USA, June 8–12, 2008

Paper 8026

5. C. W. FORSBERG, C. RENAULT, C. LeBRUN, E. MERLET-LUCOTTE, and V. IGNATIEV “Liquid Salt Applications and Molten Salt Reactors (Les applications des sels liquides et les reacteurs a sels fondues),” RGN: Revue Generale Nucleaire, No. 4, pp. 63–71, July–August 2007.

6. D. F. WILLIAMS, L. M. TOTH, and

K. T. CLARNO, Assessment of Candidate Molten Salt Coolants for the Advanced High-Temperature Reactor (AHTR), ORNL/TM-2006/12, Oak Ridge National Laboratory, Oak Ridge, Tennessee (2006).

7. C. W. FORSBERG, V. K. VARMA, and

T. W. BURGESS, “Three-Dimensional Imaging and Precision Metrology for Liquid-Salt-Cooled Reactors,” 5th International Topical Meeting on Nuclear Plant Instrumentation, Controls, and Human Machine Interface Technology, Embedded Topical in American Nuclear Society 2006 Winter Meeting, Albuquerque, New Mexico, November 12–16, 2006.

8. D. F. WILLIAMS, Assessment of Candidate

Molten Salt Coolants for the NGNP/NHI Heat-Transfer Loop, ORNL/TM-2006/69, Oak Ridge National Laboratory, Oak Ridge, Tennessee (2006).

9. P. F. PETERSON and H. ZHAO, “A Flexible

Baseline Design For the Advanced High Temperature Reactor Utilizing Metallic Reactor Internals (AHTR-MI),” 2006 International Congress on Advances in Nuclear Power Plants (ICAPP '06), Reno, NV, June 4–8, 2006, American Nuclear Society, La Grange Park, Illinois.

10. C. W. FORSBERG, D. INGERSOLL,

P. PETERSON, H. ZHAO, J. E. CALAHAN, T. TAIWO, J. A. ENNEKING, R. A. KOCHENDARFER, and P. E. MacDONALD, Refueling Options and Considerations for Liquid-Salt-Cooled Very High-Temperature Reactors, ORNL/TM-2006/92, June 2006.

11. C. W. FORSBERG, P. F. PETERSON, J. E. CAHALAN, J. A. ENNEKING, and P. MacDONALD, “Refueling Liquid-Salt-Cooled Very High-Temperature Reactors,” ICONE14-89471, CD-ROM, 14th International Conference on Nuclear Energy, Miami, Florida, July 17–20, American Society of Mechanical Engineers, 2006.

12. S. J. deZWAAN, The Liquid Salt Pebble Bed

Reactor: A New High-Temperature Nuclear Reactor, Department of Radiation, Radionuclides and Reactors, Delft University of Technology, Delft, The Netherlands (2005).

13. P. BARDET, J. Y. AN, J. T. FRANKLIN,

D. HUANG, K. LEE, M. TOULOUSE, and P. F. PETERSON, “The Pebble Recirculation Experiment (PREX) for the AHTR,” Global 2007: Advanced Nuclear Fuel Cycles and Systems, Boise, Idaho, September 9–13, 2007, American Nuclear Society, LaGrange Park, Illinois (2007).

14. A. NIQUILLE, M. FRATONI, P. BARDET,

E. GREENSPAN, AND P. F. PETERSON, “Design, Analysis, and Development of the Pebble Bed Advanced High Temperature Reactor,” International Congress on Advanced Nuclear Power Plants, Anaheim, California, June 8–15, 2008, American Nuclear Society, La Grange Park, Illinois, 2008.

15. K. O. STEIN, R. A. KOCHENDARFER, and

J. W. MADDOX, “Use of a Liquid-Salt Nuclear Reactor (LS-VHTR) to Transmute Minor Actinides,” International Congress on Advanced Nuclear Power Plants, Anaheim, California, June 8–15, 2008, American Nuclear Society, La Grange Park, Illinois, 2008.

16. C. W. FORSBERG, Designing the Advanced

High-Temperature Reactor for Low Radionuclide Releases in Beyond-Design-Basis Accidents,” International Congress on Advanced Nuclear Power Plants, Anaheim, California, June 8–15, 2008, American Nuclear Society, La Grange Park, Illinois, 2008.