Aker Solutions is a leading global provider of engineering and construction services in the energy sector. It is not typically recognised as a reactor vendor. However, based on the potential acquisition of a thorium mine, the company investigated the optimum use of thorium. This led to collaboration with Professor Carlo Rubbia, originator of the Energy Amplifier [1], in the development of the accelerator-driven thorium reactor (ADTR) as a new power reactor. (The mine initiative has since been cancelled).

Much has been published about the Energy Amplifier concept and accelerator-driven systems. However the physics adopted means the accelerator is large-scale, potentially beyond what could be achieved as a reliable industrial unit. Hence Aker Solutions commissioned a feasibility study carried out in Stockton-on-Tees, England and Geneva, Switzerland, to develop the physics, engineering and business model for a commercial 600MWe ADTR power station based on a sub-critical, thorium fuelled, lead cooled, fast reactor with a proton accelerator of proven design. As an industrial company, the challenge was to apply existing technology to create a reactor for practical and industrial use: a power station.

Perhaps the most radical result of this approach has been selection of a neutron multiplication coefficient, keffective (keff) of 0.995. (keff is the average number of neutrons from one fission that cause another fission). Rubbia’s original paper was based around a keff of 0.98, a sub-critical configuration which provides an intrinsically safe margin from criticality for significant periods over an eight to ten year core life. To drive the choice of system gain to a much higher value and hence an accelerator of less power and cost in the ADTR, a keff of 0.995 is proposed. A method of measuring keff is required so that the core condition is known as it changes over long-term operation and also to ensure precise control is maintained at any instant of reactor operation. A unique methodology for controlling reactivity in the ADTR has been developed that involves the use of fast-acting shutdown control rods, control rods and neutron detectors spread throughout the fuel assembly [2]. This work is fundamental to commercial viability of the ADTR. Without a method for assessing keff of the system, acceptable control cannot otherwise be guaranteed.

This concept study work supports the technical feasibility of a commercially viable ADTR power station. Assessment work to date indicates an ADTR power station has a cost-per-megawatt output equivalent to conventional nuclear power systems. Now Aker Solutions is interested in discussing this exciting opportunity with potential partners. It is particularly looking for financing and expert resources.

Reactor physics

The ADTR power station is a 600MWe, accelerator-driven, thorium-fuelled, lead-cooled, power-producing, fast reactor. The ADTR core consists of a series of fuel pins gathered together in hexagonal assemblies. In the initial fuel loading each fuel pin contains mixed oxide pellets composed of 15.5% plutonium (the ‘starter’) and 84.5% thorium. The starter material—Pu from spent fuel or potentially minor actinides (MA)—is consumed and Th-232 is converted to U-233 during the operation of the reactor. Eventually enough U-233 is created to provide the starter material for subsequent cycles of the reactor following reprocessing. The fuel cycle is a closed system from year one; no further fissile material is required once the first cycle is started; only fresh thorium is required.

Boron carbide control rods, consisting of 90% enriched boron-10, are used to depress the core raw reactivity to the designated sub-critical value. A particle accelerator is used to inject high energy protons into a target—in the concept design this target is also the liquid lead coolant. These protons collide with lead atoms, causing them to fragment releasing a number of neutrons (spallation). Enough additional neutrons are generated in this manner to provide a sustainable nuclear power process within the ADTR. This spallation process provides a mechanism for control of reactor power output. Since the overall power gain of the system is proportional to the accelerator current, adjustment of this parameter results in a commensurate change to reactor power. By cutting the current completely, the accelerator also provides a means of instantly reducing reactor power on shutdown.

A preliminary fault schedule was compiled from qualitative HAZOP studies and safety reviews of the concept design which shows key faults and the range of potential causes. From this summary, a preliminary assessment of the worst case consequences identified several potential accident scenarios. The HAZOP study process is iterative and a number of studies are foreseen for subsequent project stages.

The maximum credible accident determined from the analyses is a guillotine break of the pipe connecting the main vessel to the heat exchanger. This results in an immediate loss of the main heat sink and initiates shutdown of the reactor (this sequence is led by accelerator switch-off which instantaneously reduces core reactivity). Coolant continues to circulate within the reactor vessel and removes heat from the core by natural convection. Decay heat is dissipated to atmosphere again using natural convection via a passive air system. This does not result in a degenerated core or radioactive release to the environment.

Why thorium?

Thorium offers a sustainable fuel cycle, waste treatment and waste reduction. Given the large potential worldwide expansion in nuclear generation it is likely that, in the medium term, the demand for uranium will increase. Although at present the effect of this on uranium price is only marginal; in the future this is likely to change with demand-led price rises predicted. Hence the option to utilise alternative fuels such as thorium will become more attractive. Thorium reserves are currently estimated to be three to five times more abundant than uranium.

A key benefit of using thorium fuel is non-proliferation; uranium enrichment technology is not required. Additionally using thorium in a breeding cycle as a nuclear fuel also generates isotopes which are high energy gamma emitters, therefore making handling and subsequent diversion of materials to bomb-making more difficult.

Minor actinides are a feature of waste products generated from reactors fuelled with natural or low-enriched uranium. They are the most problematic wastes in irradiated fuel to handle due to their high toxicity index and potential heat generation. Thorium fuelled reactors do not create actinides since uranium (U-238) is not present. In fact the ADTR can be configured to burn actinides whilst producing power, thus reducing the long-term waste burden.

Since thorium is fertile and not directly fissionable a fissile starter material is needed to initiate the reaction. In the ADTR reactor-grade plutonium (Pu-239) is used, although other materials such as minor actinides (principally americium and curium) could also be utilised. As the reaction progresses over a period of four weeks, the fissile plutonium content is replaced by U-233 bred from thorium. The breeding reaction is:

232Th + n → 233Th (t1/2=22min) +γ → 233Pa (t1/2=27d) + β → 233U (t1/2=150,000y) + β

The ADTR core material composition changes considerably as the initial starter material is partially consumed and U-233 (amongst other isotopes) is produced from thorium breeding.

An 8-10 year self-sustained fuel cycle is possible, thus increasing availability of the system for power generation. Operational costs are reduced, since refuelling is infrequent and fuel shuffling is not required. In addition, since access to fuel is not needed for extended periods, materials safeguards can be enhanced by use of IAEA seals which could be fitted for the fuel cycle period.

In common with other nuclear systems, isotopic poisons that accumulate in the ADTR core reduce the reaction efficiency with time. It is therefore logical to consider ADTR fuels in cyclical terms involving fuel fabrication, reprocessing and separation of the valuable isotopes, removal and disposal of the unwanted components and finally recombination of reclaimed fissile U-233 with thorium as a mixed oxide fuel.

Fig. 2 shows the development of the fuel mixture over successive fuel cycles. After each fuel cycle it is assumed that reprocessing will recover the uranium and plutonium content then use this to refabricate new fuel with fresh thorium. The figure depicts 11 successive cycles with each cycle representing approximately ten years of electricity production. As can be seen, plutonium is effectively consumed and uranium content increases until an equilibrium position is reached. (The study has not concluded whether a reprocessing facility for a single or suite of reactors is best; its capacity would depend on the requirements of a client.)

The wastes for disposal contain fission fragments and unused thorium. This dramatically reduces the toxic nature of the material sent for long term disposal compared to conventional uranium based fuels.

Why accelerator-driven?

An accelerator-driven reactor offers inherent safety and load following. A particle accelerator is used to inject high energy protons into the molten lead coolant. These protons collide and cause the lead atoms to split, releasing a number of neutrons and a variety of fragments in a process known as spallation. Typically for 1GeV protons a spallation yield of approximately 30 neutrons is achieved. These neutrons interact with other nuclei—lead, uranium, thorium, plutonium or others—and will either cause fission (in the case of U-233 and Pu-239 and liberating more neutrons in the process) or be absorbed (in the case of Th-232). A small number of neutrons are lost from the system or are absorbed in other materials in the reactor. However, enough additional neutrons are generated through spallation to provide a sustainable chain reaction within the ADTR.

The accelerator complex will comprise primarily three elements—low, intermediate and high energy sections—and will operate with a beam current of 5mA. In common with all other mechanical, electrical and control systems supplied for a commercial power plant, the availability and maintainability aspects of the accelerator need to be considered. Several configurations are possible which will ensure plant operational availability, for example during planned maintenance outages and to provide ‘hot standby’ capability.

Current typical accelerator driven systems (ADS) are based around neutron multiplication coefficients, keff, of 0.95-0.98, a sub-critical configuration which provides an intrinsically safe margin from criticality for significant periods over the core life. This translates to a nominal energetic gain of at best 120, equating to a beam power of 12.5MW with particle energy of 1GeV. Although such an accelerator can be built with existing technology it would be expensive both in terms of capital cost and in power consumption during operation. An alternative approach was based on fixing the neutron multiplication coefficient based on established accelerator technology thereby limiting project commercial and technical risk.

The energetic gain, G, is defined as the thermal energy produced by the reactor divided by the energy deposited by the accelerator beam. From [1] this is:

G = 2.4*/(1-keff) = Reactor power/(proton current x proton energy)

*2.4 is an approximate value

For the ADTR a proton kinetic energy of 1GeV was chosen with beam power set at 4MW. This equates to an energetic gain of 402 to 532 at keff of 0.995. This reactivity must be maintained throughout the fuel cycle and is achieved in the ADTR by establishing a core with excess reactivity and suppressing neutron multiplication by insertion of an absorber material.

The ‘raw’ reactivity level of the fuel mixture evolves over time. Initially reactivity decreases as the thorium is converted to the intermediate Pa-233 which acts as a reactor poison. With a half life of 27 days the Pa-233 decays to fissile U-233, which then increases reactivity. Reactivity peaks at mid-cycle and starts to decline as neutron absorbing fission products increase and start to dominate (see Fig. 3). This variation of reactivity is regulated by means of adjustable neutron absorbing control rods. The raw reactivity characteristics seen in Fig. 3 have remarkable stability over the required fuel life cycle.

Since power can be cut virtually instantaneously to the accelerator, the ADTR responds extremely rapidly to any reactivity transient. Reactor power will reduce rapidly following a stop in the beam current with full shutdown achieved by use of more conventional control/shutdown rods. In theory this type of control compares very favourably to conventional reactor systems that rely solely on mechanical devices for control and shutdown rod insertion. (The final reactor design will include secondary shutdown systems, such as absorbent material injection into the accelerator beam tube, but these details are not yet finalised.) The accelerator provides a simple means of varying the power output of the reactor thus facilitating load following. A variation of 10% in current will vary reactor output by 10%—it functions as a kind of electronic power control. Operation in the fast neutron spectrum avoids the xenon poisoning problems experienced by conventional water-cooled thermal reactors.

Why lead coolant?

The big advantage of lead coolant is that it allows the creation of an atmospheric pressure system.

Operating at a slightly sub-critical level, the reaction is sustained by neutrons generated by proton spallation of molten lead, which also acts as the reactor coolant. Lead coolant is chemically non-reactive, has a negative void coefficient which enhances safe response to thermal transients, and with its large thermal mass, provides a significant heat sink in the event of any power excursion. Natural lead is cheaper than lead-bismuth and produces much less polonium. Still, Po-210 is a daughter product of lead spallation that raises the most concern. However, experience from Russian lead-bismuth cooled reactors and from high-power lead spallation targets at the Paul Scherrer Institute shows that polonium remains in the lead as a liquid eutectic.

The high boiling point of lead (~2022K; 1750°C) allows the ADTR primary circuit to operate at atmospheric pressure, reducing design demands and consequences from any failure scenario. This does not negate consideration of the various pressure systems codes for the designer. However, it does simplify aspects of the ADTR vessel design; for example, the internal pressure effects are negligible when considering the design of the upper vessel head and associated flange details.

Selection of structural materials for the ADTR requires consideration of reactions between these materials and the liquid lead coolant. This is further complicated when these structural materials are subject to sustained neutron flux. Russian institutes have investigated this over several decades achieving notable success. Recent work continues to improve knowledge in this key area. Therefore, whilst noting this is a complex area, there is confidence that suitable known and understood materials can be used.

The reactor vessel consists of three nested tubes (Fig. 4). In the innermost tube, the proton beam travels to the centre of the fuel assemblies. In the middle tube, spallation neutrons generate heat and uranium in the fuel assemblies in a liquid lead bath. Lead heated by the nuclear reaction rises to tubes near the top of the reactor (exit temperature: 550°C) that carry it to heat exchangers that transfer the heat to a water-based secondary loop that leads to a turbine. Cooled (420°C) lead returns to the top of the reactor, and then flows down the outermost tube to the base of the rector vessel, beneath the core, to which it returns through perforations in the base of the middle tube. Although preliminary calculations had considered relying on natural convection, that circulation process determined to be uneconomical to generate the required heat transmission. Instead, four axial flow pumps have been included in the system. The reactor vessel is supported by an integral ring beam, which transfers the load of the reactor system to a massive concrete support shelf. Although the middle tube faces the highest temperatures, it is a non-structural component, so there is greater freedom in material selection.

In the highly unlikely circumstances that the fuel clad should melt, any materials arising will tend to float away from the core, effectively limiting reactivity; the process is self-limiting.

Why 600MW(e)?

The approach taken to this study has been driven by the need to establish commercialisation of the ADTR. For the ADTR to be a viable business opportunity, it needs to be at the forefront of the leading Generation IV reactors to market. Hence ‘time to market’ for the first operational ADTR power station could be by approximately 2030; consistent with the commercialisation of other Generation IV reactors.

The ADTR could take a significant market share, particularly in specific markets where a nuclear infrastructure is not already established, since the infrastructure necessary for uranium systems (such as enrichment plants) is not necessary. The 600MW size of the ADTR fills a market gap between small modular systems and the Generation 3+ systems (which are in excess of 1000MW). The 600MW size and the potential to load follow using the accelerator means the reactor operation is flexible and hence could be beneficial for smaller developing countries wishing to move to nuclear power generation but with less mature grid systems not readily capable of accommodating large generating units.

Author Info:

Victoria B. Ashley, project manager; Roger Ashworth, senior physicist; John E. Earp, project director; Colin G. Fuller, engineering manager.

Aker Solutions, Phoenix House, 3 Surtees Way, Surtees Business Park, Stockton-on-Tees, TS18 3HR, United Kingdom

Related Articles
A thought for thorium
India to invest in ADS

Reactor diagrams