|
|
| Fig. 1: Illustration of a TRISO fuel particle. It is about 0.8 mm in diameter and consists of layers of carbon and SiC surrounding a fissile core. [19] (Courtesy of the DOE. Source: Wikimedia Commons) |
Generation IV nuclear reactors represent the next frontier in nuclear energy technology, promising significant advancements in safety and efficiency over their Generation III predecessors. Among these innovative designs, two reactor technologies have presented great promise of being licensed: High-Temperature Gas-Cooled Reactors (HTGRs) and Sodium Fast Reactors (SFRs). [1] With existing regulatory paradigm for commercializing new nuclear power technologies, these two reactor technolgoies show the high promise of being licensed in next decade.
High-Temperature Gas-Cooled Reactors (HTGRs) are helium cooled graphite moderated nuclear fission reactors which commonly use fully ceramic uranium dioxide (UO2) tri-structural isotropic (TRISO) particle-based fuels. [2] These reactors are distinct for their use of helium gas as a coolant, which allows them to operate at temperatures from 700°C to 950°C, far higher than the 330°C of traditional nuclear light water reactors (LWRs). [1,2] The significantly higher operating temperatures of HTGRs compared to LWRs permits greater thermal efficiency, translating to more electricity generated per unit of fuel consumed, with some proposed designs presenting thermal conversion efficiencies of up to 48%, a 1.5 factor increase compared to the 30% efficiencies recorded in LWRs. [1,2]
Every heat engine is restrained to the theoretical efficiency limit of the Carnot Efficiency. Assuming that the cold sink temperature is at room temperature at 25°C for both reactor designs, we can see how the theoretical efficiency limit grows by increasing the temperature from 330°C in LWR to 950°C in HTGRs:
| 330°C: | 1 | + | 25°C + 273° 330°C + 273° |
= | 0.506 | |
| 950°C: | 1 | + | 25°C + 273° 950°C + 273° |
= | 0.756 |
As can be seen, the efficiency theoretical limit also increases by a factor of around 1.5, highlighting the theoretical feasibility of higher temperatures achieving a 1.5 increase in efficiency. However, a theoretical possibility does not directly translate into a commercial reality.
While high efficiencies may represent an energetic gain, they do not translate into an economic gain directly. In the energy market, it is more important to consider the watts generated per dollar spent rather than solely the efficiency of the energy generated. Simply put, a gain in efficiency is only an economic gain if it increases the amount of energy generated for the same amount of dollars spent. On the other hand, if the gain in efficiency results in proportionally larger costs, the watts generated per dollar spent will go down and the efficiency gain results in an economic loss. This is particularly important when considering reactors operating at higher temperatures, such as HTGRs, as temperatures above 600°C require specialty nickel steels and much thickened metallic walls to avoid steel creep, which would occur in conventional reactor steels. [3] Given nickel's cost of $22/kg compared to chromium and steel scrap costs of $0.5/kg and $0.3/kg, respectively, it is clear that a steel alloy containing higher percentages of nickel will likely be more expensive than conventional reactor steel. [4] Beyond the high mineral costs, nickel steel (Incaloy) is extremely challenging to fabricate and weld and could significantly increase construction costs. As a result, the efficiency gain must be balanced with the extra costs associated with using high-temperature alloys. Material challenges could explain why commercial HTGRs, such as the Fort Saint Vain, have operated at the lower end of the HTGR temperature spectrum at around 777°C, achieving efficiencies of 39%, below the proposed efficiencies presented in the GIF 2022 report. [5]
Most HTGRs use UO2 TRISO coated particles as fuel. [2] These particles made up of a UO2 kernel, an inner pyrocarbon (porous carbon) buffer, a silicon carbide middle buffer layer and an outer pyrocarbon coating embedded into a graphite matrix. [2] The TRISO fuel has the unique advantage withstanding extreme temperatures. Tests at Idaho National Laboratory and Oak Ridge National Laboratory in 2013 showed that the uranium fission products remained contained inside the fuel particles after being subjected to temperatures of over 1800°C (above both normal reactor operating temperatures and postulated accident conditions of 1600°C). [6] By withstanding such high temperatures the fuel mitigates the risk of the radioactive fuel leakage, as each TRISO particle acts as a containment system withholding the fissile material inside the particle. One note should be made that containing the fissile material does not mean that the reactor is not prone to radiation leakage, as fission reactions occurring inside the reactor vessel also make other materials in the reactor radioactive that could still contaminate surrounding areas in the case of an accident.
It should also be noted that all TRISO fuel designs today, and those currently produced by Oak Ridge National Laboratory (ORNL) use HALEU (high assay low enriched uranium) feedstock. [1] Currently, only Centrus Energy has permission by the NRC to produce HALEU feedstock with a limit of 900 kg/year and an enrichment limit of 19.5%. [7] Lack of wide scale commercial use of the HALEU feedstock could provide a challenge in the deployment of power plants that rely on HALEU feedstock possibly rising costs for the plant. Nonetheless today fuel costs represent around 5% of overall plant lifetime costs, meaning even if HALEU is twice as expensive as regular low enriched uranium, the impacts will be small compared to the other sources of costs in nuclear plants. [1] However, even if the costs do not impede deployability of TRISO fuel, it remains to be seen how public opinion will respond to the fuel's introduction. Its high safety characteristics stand to improve its view in the public eyes but the use of HALEU may be questioned with worries of highly enriched fuel being used in commercial reactors.
Today, GIF member countries are forming their TRISO fuel into two main structures: either 60 mm diameter pebbles or cylindrical compacts embedded in hexagonal fuel blocks. [2] The different fuel designs result in different reactor designs and capabilities, with the pebbles giving rise to pebble bed reactors and the cylindrical prisms resulting prismatic reactors.
The pebble-bed design was originally developed in Germany by the Jülich Research Institute and utilizes spherical fuel elements, known as pebbles, which are composed of thousands of TRISO fuel particles organized in a prism design. [8] The AVR, the original pebble bed reator developed by Jülich, initially operated with Bi-structural isotropic (BISO) coated carbide fuel elements yet in 1982 they replaced the BISO fuel with TRISO fuel, achieving extremely low release of fission products at temperatures above 950°C. [8] The AVR system showcased the unique advantage of the pebble bed design in comparison to the prismatic design as it allowed for online refueling, reducing shutdown times for the reactor and reducing accident risks resulting from powering down and powering up reactors. Current capacity rates for LWRs average around 85%, mostly due to refueling needed every 18 months. [1] In 1976 the AVR reactor achieved a 91.9% capacity factor showcasing the significant advancements that can achieved by introducing the pebble bed design. [8] Commercial designs of this reactor would not experience planned downtimes for experimental preparations, testing modifications and repairs, likely resulting in noticeably higher capacity factors and a significant gain in watts generated per dollar spent, especially given the low fuel cost. [8] Several institutes and companies have looked to re-invent this initial design. China National Nuclear Corporation and Tsinghua University have deployed a commercial version of this reactor in Shindao Bay, Shandong Province with the information being widely reported in Chinese and Western press, yet the is little public scientific documentation on the project as of this writing. Other designs under consideratin include a one proposed by the American start-up X-energy and MIT. [1,9]
The MIT pebble bed reactor plant design (MPBR) has a large body of openly available research on its designs which can be best cited and evaluated. [9] The design was largely based on the decommissioned pebble bed modular reactor (PBMR) project initiated in South Africa by Eskom. Their design aimed for a modular plant and reactor with 250 MWth power (or 120 MWe power), seeking to achieve 45% thermal efficiency, using 360,000 60 mm fuel pebbles weighing 7 g each, with a fuel enrichment of 8%. Given the AVR design's success in operating at temperatures above 950°C, the MPBR would aim to operate at 900°C with He at 80 bar, using a total of 6 control rods. [9,10] The design would also operate with an intermediate heat exchanger, which adds costs in design but allows for the use of more conventional systems for the power conversion cycle. [9] The MIT design also added constraints for modularity - meaning it worked with six heat exchangers/recuperators, which could likely increase costs - yet was also responsible for an increased efficiency of > 45%. [9]
While the MIT plans detailed the advantages of their project, little was discussed of the material challenges that could be encountered by operating at such high temperatures. As mentioned previously high temperatures can result in steel creep and result in the need for speciality nickel steel alloys. In the MIT design, which would operate at temperatures exceeding the previously studied 700°C it is quite likely that an even more advanced steel alloy would have to be introduced, certainly exacerbating costs. Beyond that, the proposed efficiencies should be evaluated in light of existing operational evidence of HTGRs. The AVR design, the source of inspiration for the PBMR and consequently for the MPBR, operated for 21 years and only ever achieved thermal efficiencies of 32% (generating 15 MWe from 46 MWth), far below the proposed efficiencies in the MIT design. [10] Nonetheless, it is worth adding that the the AVR reactor was not focused on energy generation, serving rather as an experimental prototype project to evaluate the safety and operability of the pebble bed technology, possibly allowing for efficiency advancements in commercial designs. [10]
As we look ahead into the future most proposed pebble bed designs, such as the X-energy design, have stayed within the experience margins of the AVR facilitating licensing procedures. [1,10]
The prismatic block design uses hexagonal graphite blocks as the moderator, with channels drilled through for the fuel and coolant. [1] These graphite blocks also contain TRISO particles, ensuring high levels of fuel integrity and containment. [1] Similar to the pebble bed reactor, the prismatic reactors also experienced testing successes, with the Fort Saint Vain Reactor (FSV) in the USA representing a notable commercial deployment of technology, operating from 1974 to 1989. [5] The plant consisted of a core of 1482 hexagonal fuel elements stacked in 6 layers with an initial core containing 774 kg of 93.5% enriched uranium and 15,905 kg of thorium. [5] The fuel consisted of TRISO-coated microspheres of uranium/thorium dicarbide. [5] As with most HTGRs, helium coolant was used with a core outlet temperature of 777°C. [11] The high outlet temperatures meant that out of a total of 842 MWth of thermal energy produced a net 330 MWe of electrical power was generated translating to a 39.1% efficiency. [5] These efficiencies far exceed those of the operational pebble bed design at AVR, likely given the fact that this reactor was a commercial design aimed to achieve profitability and thus sought to make the most of its high temperature output, different from the AVR design that served as a research experiment. Interestingly the recorded efficiencies are below the efficiencies aimed for in the MIT design, raising possible questions on whether the MIT efficiencies can be achieved in a commercial design considering the added costs of replacing traditional reactor steel with Incaloy in the coolant ducts.
The biggest difference between the prismatic block and pebble bed design is that the prismatic block does not allow for active refueling. During the 15 years of operation Fort Saint Vain experienced an initial loading of the fuel and three additional refuellings, meaning that the fuel cycle averaged at 3.75 years (45 months). [5] While this outage period is certainly inferior to the pebble bed design it still presents a significant improvement over existing light water reactors that have a fuel cycle time of 18 months. This advantage is likely due to the use of HALEU fuel which provides more fissile fuel for burn up so it must be tempered with the added cost of innovative fuel sourcing, as detailed above.
The FSV provides us an additional insight on HTGRs that pebble bed reactors still have not been able to clarify. As a commercial reactor, the facility operated to sell electricity and make a profit, yet despite operating for 15 years it was shut down, facing a far shorter lifetime than traditional LWRs. During a review conducted by the ORNL between 1981 and 1989, 279 events were reported of some type of operational failure or flaw detected in the system, providing us some insight into the source of the shutdown of the reactor. [11] Out of these events, the biggest challenge lay with moisture intrusion where small amounts of moisture degraded both the control rod drive and the reserve shutdown systems, leading to six control rod pairs failing to scram during an event on June 1984. [11] The extent of the damage caused by the moisture intrusion was not well understood until in 1989 when the plant was shut down to repair a stuck control rod pair. During the shutdown, numerous cracks were discovered in several steam generator (SG) main steam ringheaders. The required repairs were determined by the PSC board of directors to be too extensive to justify continued operation, so they decided to permanently terminate nuclear operations on August 29, 1989. [11] The moisture challenges resulting from the new design certainly need to be considered for any company seeking to reintroduce this technology.
Sodium Fast Reactors employ liquid sodium as a coolant, allowing for operations at higher temperatures without the high pressures associated with traditional water-cooled reactors. [1] SFRs are breeder reactors, meaning they work to generate more fissile material than they consume (with breeding ratios greater than 1). [2] The technology operates without a moderator, resulting in a very high-power density. Outlet temperatures for SFRs range between 500°C to 550°C at atmospheric pressures. [2] At these temperatures austenitic and ferritic stainless steel structural materials can be used, especially given that they are compatible with sodium. [2] This stands as a advantage to HTGRs that require Incaloy due to the higher outlet temperatures, allowing for cheaper coolant ducts. [2] However, as was experienced with experimental and commercial SFRs, the greatest challenge for SFRs lies with the use of sodium as a coolant. Sodium buildup and high flammability result in significant maintenance costs and down time for the reactor reducing its capacity factor and consequently significantly reducing its watts per dollar spent. [12-15]
Amongst GIF member nations SFRs are commonly classified into two main configurations: loop and pool designs, as outlined in Fig. 2. The loop design has the primary coolant leaving the reactor vessel to the intermediate heat exchanger (IHX) located in the containment area outside the vessel, allowing for easier maintenance in the heat exchanger. [16] The pool design has the primary coolant kept within the reactor vessel, which also encompasses the IHX, introducing challenges to maintenance but reducing the impact of a primary pipe break or leak. [16]
 |
| Fig. 2: Pool vs Loop Reactor Designs. [16] (Source: Wikimedia Commons) |
A widely studied and actively operated SFR pool designed reactor was the EBR-II, from the Argonne National Laboratory in the USA. First achieving criticality in 1964, the EBR-II was a 62.5 MWth (20 MWe) reactor operating with a primary and secondary coolant structure, where the primary coolant outlet temperature exited at 472.8°C, while the secondary coolant exited at 466.7°C. [12] The lower temperatures and the reactor's primary design as a testing facility resulted in its lower thermal efficiency, of 32%, compared to HTGR proposed designs and the operational Fort Saint Vain plant. The reactor was operational until 1994 when it was shut down.
The EBR-II also highlighted a recurring challenge with SFRs regarding their capacity factor. From 1976 to 1982, the EBR-II capacity factor averaged 73.7%, yet in prior 11 years operational and maintenance challenges kept capacity factors below 60% or even below 50%. [13] The majority of the operational challenges came from buildup of sodium and sodium oxide in cover-gas spaces of primary pumps and fuel handling components, causing binding and sodium buildup in the clearances. [12] Corrective actions had to be taken by cleaning the component, enlarging clearances where feasible and improving sodium drainage. [12] By 1976, Argonne National Laboratory seemed to have learnt well how to operate the SFR explaining its higher capacity factory, suggesting a possible breakthrough in the technology. [13] Overall, the reactor and its design proved the relative safety and operationality of SFRs, with the higher capacity factors reflecting a maturation in operational skills in dealing with motel metal coolants. Nonetheless, even with the improved capacity factors, the plant was generating energy at far lower rates compared to regular LWR (85%) and far below pebble bed reactor accomplishments at the AVR, reaching a capacity factor of 91.9%.
As an experimental prototype the EBR-II reactor provides limited insights into the true challenges and capabilities of SFRs. To best understand the future prospects of this technology it would be best to evaluate a commercially operational plant, such as was done in the case of prismatic block HTGRs with the FSV reactor. For SFRs the key commercial case study is the Superphénix in France. In light of the growing worry of the limited availability and high costs of uranium in the 1970s France launched Superphénix, its largest breeder reactor aimed to maximize the country's use of uranium fuel. [17] The Superphénix was a pool design SFR planned to generate 1240 MW of electrical energy from 3000 MW of thermal energy, translating to an expected efficiency of 41.5%. These high efficiencies were a result of the reactor's high outlet temperature of 545°C. [17] The plant had an expected breeding ratio of 1.24 with refueling operations scheduled for every 12 months. [17]
True to its plans, the Superphénix achieved 45.3% efficiency averaging around 40% during its operational time. [14] However, while the efficiency gain was achieved the facility faced other problems that resulted in its closure. At a power rating of 1240 MW and operating from 1986 to 1997, the plant could have generated at most 1240 MW × 24 hours/day × 365 days of operation = 108,624,000 MWh, or 108,624 GWh of electricity. Superphénix produced a grand total of 7494.72 GWH of electricity before it was decommissioned in early 1997. [18] This translates to a capacity factor of 6.9%, exponentially smaller than the 85% achieved in LWRs and the 91.8% achieved by the AVR. Sodium leakages, sodium fires and general challenges in dealing with the sodium coolant ultimately meant that the reactor that should have been operational 85% of the time was operational only 6.9% of the time. [14] Different from the experimental reactor, the Superphénix showed us that at a larger commercial scale SFRs have a significant challenge in dealing with liquid sodium as a coolant.
The challenge of liquid sodium as a coolant is not constrained solely to mega scale reactors, such as the Superphénix. The UK's the Prototype Fast Reactor (PFR) at Dounreay, around 1/5th the size of the Superphénix but still 10 times larger than the EBR-II, was also marked by a low lifetime load capacity factor, recorded at 26.9%. [15] The PFR was a 630 MWth (235 MWe) reactor, with a 37.3% thermal efficiency given a 562°C outlet core coolant temperature. [15] The reactor operated with a 25% enriched fuel rod with high burn up rates of 15% of the heavy nuclei. [15] While achieving higher efficiencies than traditional LWRs the PFR remained constrained from commercial applications due to its low capacity factors, similar to the Superphénix.
Looking forward HTGRs and SFRs present nuanced advancements and distinct challenges which paint a comprehensive picture of their potential roles in the nuclear energy sector. Both technologies were commercially tested with both ultimately facing a shutdown after a decade or more of operation. In the HTGR end, the pebble bed design remains to be evaluated at a commercial setting, with the Chinese HTR-PM hopefully providing commercial insight in the next decade or two. The prismatic block design was commercially tested and operated well, but design flaws ultimately shutdown the project prematurely. Future endeavors will need to address the issue of moisture intrusions that came with the adoption of a prototype helium cooled reactor design. [11] These challenges will also be an important consideration for pebble bed designs that also work with a helium coolant. In the ORNL report, the laboratory highlighted that many of these design flaws could be addressed in future HTGR designs, highlighting the theoretical feasibility to out-engineer these challenges. [11]
At the SFR end, commercial experiences were more extensive yet revealed more significant challenges to future commercial viability, mostly due to the maintenance challenges of a sodium coolant. The significantly low capacity factors of the Superphénix and the PFR suggest that at substantially sized reactors dealing with large volumes of a highly flammable high temperature liquid metallic coolants can pose significant operations and maintenance challenges. [14,15] As new companies aim to re-introduce SFRs they must evaluate in depth the obstacles faced in the Superphénix and best engineer their solution around the problem. The late stage operational success of the EBR-II in achieving higher capacity factors suggest that experience can be acquired in managing a sodium coolant, especially with smaller scale reactors. [13]
Beyond the operational challenges and successes of each technology, some technical considerations are of important consideration when looking into how both technologies will succeed in the future. A key factor to consider is the operational temperature of each reactor. HTGRs, with their operational temperatures exceeding 750°C, promise a 50% increase in thermal conversion efficiencies yet at the cost of using specialty nickel steels to withstand high temperatures, potentially diminishing the economic advantage provided by the increased efficiency. In contrast, SFRs promise more moderate gains in efficiencies of around 15%, but benefit from the use of cheaper/more usual materials such as austenitic and ferritic stainless steel, which are compatible with the sodium coolant used in these reactors. The comparative economic advantage of each reactor therefore will be directly related to material advancements and depend on how much cheaper Incalloy metals will become in the future. If the costs of materials face a significant decline, or alternative alloys are developed, HTGRs may face an inherent advantage due to their higher efficiency.
The safety and fuel considerations for both reactor types underscore the importance of technological innovation in addressing the inherent challenges of nuclear power. HTGRs benefit from the robust safety features of TRISO fuel, while SFRs must contend with the reactivity of sodium coolant. The development trajectory of both technologies will likely be influenced by advancements in materials science, fuel availability and economics, and evolving regulatory standards. Nonetheless, it is critical to note that both reactor designs have abandoned the water "fail safe" present in the LWRs which in at least two cases has adverted a major catastrophic nuclear accidents, minimizing the extent of the accident. Notably the worst nuclear accident recorded in human history, Chernobyl, occurred in a power plant that operated with graphite moderation.
As the nuclear market evolves, the market competitiveness of HTGRs and SFRs will be determined by their ability to offer efficient, cost-effective, and safe nuclear power solutions. The balance between operational efficiency and material costs, alongside safety and regulatory compliance, will dictate their viability. The path forward for nuclear energy, characterized by these innovative reactor designs, remains grounded in the universal trade-offs of nuclear power, including waste management and environmental considerations. Yet, as these technologies vie for a place in the energy sector, their success will hinge on addressing these challenges, with the ultimate goal of providing sustainable and safe nuclear energy options.
© Freddy Rabbat-Neto. The author warrants that the work is the author's own and that Stanford University provided no input other than typesetting and referencing guidelines. The author grants permission to copy, distribute and display this work in unaltered form, with attribution to the author, for noncommercial purposes only. All other rights, including commercial rights, are reserved to the author.
[1] "The Future of Nuclear Energy in a Carbon-Constrained World," Massachusetts Institute of Technology, 2023.
[2] "Annual Report 2022," Generation IV International Forum, 2022.
[3] P. Mayr et al., "Microstructural Features, Mechanical Properties and High Temperature Failures of Ferritic to Ferritic Dissimilar Welds," Int. Mater. Rev. 64, 1 (2018).
[4] "Mineral Commodity Summaries 2024," U.S. Geological Survey, 2024.
[5] L. L.Taylor, "Fort Saint Vrain HTGR (Th/U carbide) Fuel Characteristics for Disposal Criticality Analysis," U.S. Department of Energy, DOE/SNF/REP-060, January 2001.
[6] "TRISO Fuel Developent Progresses at INL, ORNL," Nuclear News 56, No. 9, 78 (November 2013).
[7] K. L. Fitch, "Submittal of Reference Material in Support of License Amendment Request for the American Centrifuge Plant," Centrus Energy, December 2022, Encl. 2.
[8] E. Ziermann and G. Ivens, "Abschlußbericht für den Leistungsbetrieb des AVR-Versuchskernkraftwerkes," Verlag des Forschungszentrums Jülich, 3448 (1997) ["Final Report on the Power Operation of the AVR Experimental Nuclear Power Station," U.S. Nuclear Regulatory Commission translation 3638].
[9] A. D. Kadak, "MIT Pebble Bed Reactor Project," Nucl. Eng. Technol. 39, 95 (2007).
[10] P. Pohl, "The Importance of the AVR Pebble-Bed Reactor For the Future of Nuclear Power," AVR GmbH, August 2007.
[11] D. A. Copinger and D. L. Moses, "Fort Saint Vrain Gas Cooled Reactor Operational Experience," Oak Ridge National Laboratory, ORNL/TM-2003/223, September 2003.
[12] H. W. Buschman, K. J. Longua, and W. H. Penney, Operating Experience of the EBR-II Intermediate Heat Exchanger and the Steam Generator System, Argonne National Laboratory, CONF-830905-12, September 1983.
[13] W. H. Perry et al., "Seventeen Years of LMFBR Experience: Experimental Breeder Reactor II (EBR-II)," Argonne National Laboratory, CONF-820465-2, April 1982.
[14] J.-F. Sauvage, "Phénix - 30 Years of History: the Heart of a Reactor," Électricité de France, July 2004.
[15] S. E. Jensen and P. L. Ølgaard, "Description of the Prototype Fast Reactor at Dounreay," Risø National Laboratory, NKS/RAK-2(95)TR-C1, December 1995.
[16] H. Ohshima and S. Kubo, "Sodium-cooled Fast Reactors (SFRs)," in Handbook of Generation IV Nuclear Reactors, 2nd Ed., ed by Igor L. Pioro (Woodhead Publishing, 2023), pp. 173-194.
[17] G. A. Vendryes, "Superphénix: A Full-Scale Breeder Reactor," Sci. Am. 236, No. 3, 26 (March 1977).
[18] S. Abdul-Kafi, "The Superphnix Fast-Breeder Reactor," Pbysics 241, Stanford University, Winter 2011.
[19] B. E. Wells, N. R. Phillips and K. J. Geelhood, "TRISO Fuel: Properties and Failure Modes," Pacific Northwest National Laboratory, PNNL-31427, June 2021.
