The sodium-cooled fast reactor (SFR) uses a closed fuel cycle. Because it is a closed cycle system, its primary benefits are effective actinide management to minimize waste toxicity and optimal use of fuel through recycling. SFR is a relatively well-developed technology. Developments include the Phenix end-of — life tests planned in France, the restart of the Monju reactor in Japan, the lifetime extension of BN-600 and start-up of BN-800 in Russia, and the start-up of the China Experimental Fast Reactor (CEFR). The overall GIF plan for the SFR within Generation IV (Fig. 13.15) is based on optimization of design and operating parameters within the next few years, the building and testing of SFR plants by 2015 and commercial operation by about 2020 (Lineberry and Allen, 2002). There are three possible layout options for the reactor unit:

1 a small size (50 to 150 MWe) modular-type reactor with U-Pu-MA-Zr metal alloy fuel, supported by a fuel cycle based on pyrometallurgical processing in facilities integrated with the reactor

2 an intermediate-to-large size (300 to 1500 MWe) pool-type reactor with oxide or metal fuel

3 a large size (600 to 1500 MWe) loop-type reactor with mixed U-Pu oxide fuel and potentially MAs, supported by a fuel cycle based upon advanced aqueous processing at a central location serving a number of reactors

Figures 13.16 and 13.17 show two of the larger plant designs. Fluids that offer high performance in terms of thermal efficiency, safety and reliability include water/steam, supercritical carbon dioxide and, more recently, molten salt. Liquid sodium is used as a coolant in SFRs, enabling high power density with a low coolant volume fraction. Because sodium reacts with air and water, a sealed coolant system is required. The oxygen-free environment has the added benefit of preventing corrosion. The outlet temperature range is 500-550 °C, which is suitable for existing materials developed and tested in earlier fast reactor programs. The SFR closed fuel cycle enables regeneration of fissile fuel and facilitates management of high-level waste, particularly Pu and MAs. Its fast neutron spectrum will extend available U resources in comparison with thermal

AHX

chimney

PDRC

piping

13.16 Pool-type KALIMER (600 MWe) SFR.

reactors. Fast spectrum reactors like SFRs have the capacity to exploit almost all of the energy present in U versus the 1% utilized in conventional thermal spectrum systems (Lineberry and Allen, 2002). Fast reactors also offer a unique solution to the problem of actinide management because they operate with high-energy neutrons that are more effective at fissioning transuranic actinides. Key properties relating to actinide management include:

• consumption of TRUs in a closed fuel cycle, thereby lowering the radiotoxicity and heat load, and so facilitating waste disposal and geologic isolation

• more efficient use of U resources through multi-recycling and better management of fissile materials

• high level of safety through inherent and passive means that accommodate transients and bounding events with significant safety margins

Issues that need further research include improving safety by the reduction or even elimination of routes that can lead to severe hypothetical core disruptive accidents (CDAs). Another issue is cost. Recent studies have estimated that the capital cost of current designs may be 25% greater than conventional LWRs. Capital costs can be reduced through a combination of simpler configuration, advanced fuels and materials, and refined safety systems. There is a need to improve the management of waste generation, including improved thermal efficiency, better utilization of fuel resources, and development of superior waste forms for the SFR closed fuel cycle. Reducing the amount of waste generated from operations, maintenance and decommissioning is also an important goal, as is limiting environmental discharges and improving proliferation resistance (Vezzoni, 2011; Sagayama, 2011). There is also a need to develop ion exchange systems for reducing the volume of ceramic material required (Lineberry and Allen, 2002).

Overall radioactivity naturally also reflects the isotopic composition of spent fuel. Spent fuel radiation is from three sources of radionuclides:

1 transuranic actinides (the majority of a emitters are in this category and they are also a source of neutrons)

2 fission products (predominantly p-у emitters)

3 activation products from materials of fuel elements other than U pellets (also predominantly p-у emitters)

The total and key isotope radioactivity against spent fuel cooling time is shown in Fig. 15.8 for two fuel burnups.

15.8 Effect of cooling time on radioactivity levels of uranium oxide fuel at burnups of 45GWd/tU and 100GWd/tU.

For environmental protection and economic reasons, process water and reagents are recycled where possible. Thus, water vapour from evaporation is condensed and nitrous vapour is directed to an absorption tower to be recombined into nitric acid.

In accordance with regulations governing radiation protection (see Chapter 2), effluent containing very low levels of activity can be released without treatment to the sea. Where the activity is higher the radionuclides are removed by co-precipitation. The resulting solid wastes are then immobilized. Wastes of relatively low specific activity that do not contain alpha emitters may be incorporated in bitumen.

Higher activity liquid wastes are concentrated by a set of evaporators. These include nitric effluents, those having a significant content of a emitters, tritiated effluents (before vitrification) and non-tritiated refined products from purification of uranium and plutonium.

Raffinates — the concentrated solution of fission products and minor actinides in nitric acid — are stored in permanently cooled and ventilated pulsed tanks prior to vitrification.

P&T could provide much needed flexibility ahead of possible regulatory changes and the difficult issue of accounting for uncertainties in the long-term performance of deep geological repositories.

In fact, the management of uncertainty is an essential feature of the safety case of a geological repository.53 The role of P&T can be seen as a measure to mitigate the importance of the uncertainty, which is inherent to the very long-term nature of the radioactivity. This is achieved essentially by the reduction of the source term.

To ensure acceptable transport safety within their own borders, countries adopt lawfully mandated minimum transport safety regulations. Based on experience in the 1940s, it was concluded that if the various countries’ regulations are not based on a consistent approach, a confusing and difficult to apply set of varying requirements can result. Most countries in the world therefore have come to rely on the United Nations’ organizations, working with them in developing a suitable set of safety requirements that can be applied by all countries to all modes of transport and for all classes of dangerous goods. The national regulations can either use these requirements directly (through verbatim incorporation or incorporation by reference) or indirectly (e. g. by rewriting them to fit into the structure of their national regulations).

The UN Committee of Experts, the UN Model Regulations and the IAEA

A committee of experts was appointed by the United Nations Economic and Social Council (ECOSOC) in 1953 to develop a universal system of recommendations for the transport of dangerous goods. These recommendations were focused on reducing risks and costs in the expanding international trade in dangerous goods, with the intent that they could also be adopted for domestic purposes. ECOSOC appointed the Committee of Experts on the Transport of Dangerous Goods (CETDG) to develop the basic approach that would be applied to the packaging and transport of all dangerous goods. CETDG’s report was to take the form of recommendations, and it would be up to the national and international bodies responsible for regulating the carriage of dangerous goods to decide the extent to which these United Nations recommendations should be given the force of law.

In December 1994 the CETDG decided that its recommendations were now sufficiently complete to be recast as Model Regulations that were addressed to all governments and international organizations concerned with the development of national and international regulations for the transport of dangerous goods. This resulted in restructuring the recommendations so that they could be used directly. In July 1995, in Resolution 1995/5, ECOSOC agreed with this approach and invited all interested parties ‘when developing or updating appropriate codes and regulations, to take full account of the recommendations, including the structure and format of such codes and regulations’ (IAEA, 1998). The UN Model

Regulations, which have been periodically updated, provide a complete set of requirements for the transport of all dangerous goods. While they do not have the force of law, they are widely used as the basis for national and international regulations.

For Class 7 (radioactive material) dangerous goods, the UN Model Regulations, issued by the UN Economic Commission for Europe (ECE), are derived from the IAEA Regulations for the Safe Transport of Radioactive Material as discussed further in Section 19.2.3. The content, format and structure of the UN Model Regulations have been closely coordinated with the IAEA, ICAO, IMO and the regional land-transport modal authorities (for rail, road and inland waterway transport) to facilitate easy integration into the binding regulations of all of these bodies.

Equation 5.2 is the formulation for LCOE provided by PCGE. The top row represents the summation of the various costs (apart from ‘O&M’, which stands for operation and maintenance, the notation should be obvious), discounted at a rate r for the year in which they occur. The bottom row expresses the total amount of electricity generated, again, discounting this according to the year in which it was produced.

^ ((Investmentt + О & Mt + Fuelt + Carbonf + Decommissioningf)(l + r) ‘)

[5.2]

(Electricityt (1 + r) )

Computations of the LCOE can be greatly simplified by making a few assumptions, namely, that:

• the investment cost occurs in year zero and includes interest paid on capital (assumed for simplicity to be at the discount rate) during the period ofconstruction

• the discount rate (r) is constant with time

• costs that occur during the operational period (O&M, fuel, carbon and decommissioning) have constant yearly values

• the same amount of electricity is generated every year.

The conceptual scheme represented by these assumptions is shown in Fig. 5.2. Under these conditions it can be shown that

[5.3]

where,

CAP is the capital cost component of the LCOE;

DECOM is the decommissioning cost component of the LCOE, explained further below;

OPER is the operating and maintenance cost per unit of electricity generated;

FUEL and CARBON are the costs of fuel and carbon per unit of electricity generated;

nth is net thermal efficiency of the generating plant (which makes allowance for electricity consumed by the plant);

INVEST is the overnight cost (capital costs plus contingency, per electrical output of the plant calculated as though all the costs were incurred overnight);

Availty is availability, the ratio between the actual output and the theoretical maximum;

c is the length of the construction period (years);

f is the cost of decommissioning as a fraction of the overnight cost;

n is the length of the operational period (years);

I is the normalised financing cost i. e. the multiplier that must be applied to the overnight cost because of the delay (caused by the construction time) in creating the asset;

R is the capital recovery factor;

r is the discount rate; and

8760 is the number of hours in a year.

Total decommissioning costs are assumed to equal a fixed fraction (f) of the overnight cost. These costs are met by paying a constant annual amount (DECOM) into an accumulating decommissioning fund throughout the period of operation. The fund is assumed to earn interest at a rate, i, that is lower than the discount rate. This allows decommissioning management to be treated as an operating cost and answers the argument that the application of discount rates over long periods does not respect the principle of intergenerational equity.2 This approach is equivalent to that used for waste
management and disposal where the costs are included in the price of the fuel. In this case

[5.4]

n x Availty x 8760 (1 + z’)((l + i)n -1)

An advantage of Eq. 5.3 is that it provides a clear demonstration that the LCOE scales linearly with decommissioning, maintenance, fuel and carbon costs. The influence of net thermal efficiency and availability also become apparent. Cost discounting is represented by two terms: the first of which (I) represents financing costs and the other (R) is the capital recovery factor. This can be regarded as the annual payment that must be made, given n and r, to raise a capital sum of unity. When r = 0, R reduces to 1/n, which shows that the practice of amortising the capital cost over the lifetime of the plant, which was common practice until the 1970s, is equivalent to a discount rate of zero, an assumption that greatly favours high capital cost plant.

Uranium mining has a long and interesting history in Africa. Significant quantities of the mineral have previously come out of the Congo and Gabon. Today uranium is mined in Namibia, Niger, South Africa and Malawi.

The DR Congo, or Belgian Congo as it was then known, provided much of the uranium for the Manhattan Project in the early 1940s particularly from the Shinkolobwe mine in Katanga. There was some uranium mining subsequently by Union Miniere, to independence in 1960, when the shafts were sealed and guarded. About 25 000 tU was produced in the two decades until then.

In Gabon, the Mounana uranium deposits were discovered in 1956 by French Atomic Energy Commission (CEA) geologists and were mined from 1960 to 1999, producing nearly 28 000 tonnes of uranium from underground and open pit mining. The best known of these deposits is Oklo, discovered in 1968, which is famous for its fossil nuclear reactors, where the natural conditions about two billion years ago created at least 17 self-sustaining nuclear reactors in the wet sandstone orebody.

Namibia has two large uranium mines capable of producing 10% of world output. Rio Tinto operates the low-grade Rossing deposit, 65 km inland from Swakopmund. Rossing Uranium Ltd was formed in 1970 (now 68.6% Rio Tinto) and the company has mined the deposit from 1976 as a large-scale open pit in very hard rock. Rossing produced 3083 tU in 2010, making it the third largest uranium mine in the world. Langer Heinrich is 50 km south-south-east of Rossing and 80 km from the coast. It is being mined by Paladin Energy Ltd, producing about 1350 tU/yr.

There are several promising developments in Namibia, including Extract Resources’ Husab project on the Rossing South deposit, which promises to become one of the world’s largest uranium mines. Early in 2012 it was taken over by a subsiduary of China Guangdong Nuclear Power.

Niger has two significant long-running mines, Akouta and Arlette, supplying 6% of the world’s mined uranium. The Societe des Mines de l’Air (SOMAIR) started production from the Arlette deposit in 1971, by open cut mining. It produces about 1700 tU/yr. The Compagnie Miniere d’Akouta (COMINAK) started production from the Akouta deposit in the 1970s. This is an underground operation at a depth of about 250 metres. Production is about 1400 tU/yr.

In South Africa, uranium production has generally been a by-product of gold or copper mining. In 1951, a company was formed to exploit the uranium-rich slurries from gold mining and in 1998 this became a subsidiary of AngloGold Ltd. It produces about 500 tU/yr from material trucked in from various gold mines and from Palabora copper mine.

In Malawi, Paladin Energy has developed the Kayelekera uranium mine where production is expected to ramp up to 1460 tU/yr about mid 2012.

In the Central African Republic, Areva is developing the Bakouma project, and is ramping up production from open pit mining to 1200 tU/yr.

In Zambia, Equinox Minerals is developing the Lumwana project, which is primarily a copper mine with discrete uranium ore. Uranium ore is being stockpiled, but there is no treatment plant yet.

Reprocessing

It must first be remembered that, although a once-through cycle is an option, the use of thorium-based fuels generally assumes reprocessing of spent fuel to achieve full potential of the Th/U-233 cycle. The recovered U-233 would then be used as seed material for another cycle. Reprocessed thorium would be recycled to save natural thorium resources, since, once again, thorium is only a fertile material. Therefore, it is clear that to take full advantage of the thorium fuel cycle, it is highly desirable to retrieve the fissile U-233 recovered by reprocessing thorium spent fuel and to recycle as much as possible.

Starting in the late 1940s, the US reprocessed approximately 900 metric tons of irradiated thorium fuels to recover about 1500 kg of U-233. Other countries also recovered U-233 from thorium-based fuels. India has done so recently.

Early experience with commercial reprocessing of thorium-based fuel was limited in the US as was reprocessing in general at that time. The only US reprocessing facility to ever operate on a commercial basis was the Nuclear Fuel Services (NFS), West Valley, NY Plant. It was permanently shut down in 1972 after six years of operation. The initial core from Indian Point 1 was fabricated with HEU/thorium fuel, and it was reprocessed at the NFS Plant in late 1968. The amount of heavy metal contained in the fuel was 16 Mt. Approximately 1.1 Mt of enriched uranium containing 7 wt% U-233, 58 wt% U-235 and other uranium isotopes was recovered, and then shipped to Oak Ridge National Laboratory. It was stored there for over 15 years in liquid form, and finally processed to produce a stable oxide form. A potential stabilization process for other U-233 bearing materials was therefore demonstrated.910

Reprocessing of thorium-based fuel is somewhat more challenging than that of uranium-based fuels mainly because the dissolution of thorium metal and thorium oxide is not as straightforward as with uranium. Developed by ORNL, THOREX is a hydrometallurgic process, a derivative of the Purex process, to recover thorium and uranium from thorium-based fuel (reference 3, section 6.2). The mechanical head-end steps are similar to those of uranium-based fuels (for they are similar fuels). However, the dissolution of irradiated thorium-based fuel is slow in nitric acid. Therefore a small amount of hydrofluoric acid must be added as a catalyst to improve the dissolution process. The presence of fluoride ions causes corrosion of the stainless steel equipment (such as dissolver tanks) since fluoride has aggressive chemical properties. Consequently, appropriate buffering agents to prevent corrosion are generally needed, which complicates the design of the equipment and increases the overall reprocessing cost. Aluminium nitrate can also be added to the aqueous dissolver solution to reduce corrosion. A downside of this is that the aluminium nitrate passes through the plant to be added to the fission products and thereby increases waste production. Because of this and other factors it is expected that the THOREX process will generate 50-70% more vitrified waste than PUREX.3 Other differences from PUREX arise from the presence of sulphates, phosphates and fluorides in the reprocessing plant raffinates. These could result in considerable corrosion issues in the process of vitrifying waste materials, where high temperatures are required.

Another issue in the back-end of the thorium fuel cycle, already mentioned in Section 8.3.2, arises from the rather long half-life of Pa-233 (27 days), the precursor to U-233. As a matter of fact, the increase in reactivity resulting from the transformation of Pa-233 into U-233 after reactor shutdown must not only be taken into account in the design and operation of the reactor but also in the design of handling and storage facilities needed prior to reprocessing and dedicated to spent fuel. Practically, the cooling time before reprocessing must last at least nine months (10 times the Pa-233 half-life) or more to allow virtually complete decay of Pa-233 to U-233 (such cooling time may also be needed for other reasons, such as because of decay heat). In the THOREX process, the remaining Pa-233 is passed into the fission product waste, as would Pa-231, which is an alpha-emitting isotope in the thorium burn-up chain (produced by the (n,2n) reaction on Th-232). Studies were carried out in the 1960s to develop a process for extraction of protactinium from a nitric solution but no simple solution was found at the time. With a half-life of 30 000 years, its radiological impact could be significant for the long-term safety of disposal.

Non-aqueous processes, alternatives to THOREX, were also studied in the past, like processes of volatilization of fluorides or electrorefining, but they did not lead to any convincing result.

Finally, it should be noted that reprocessing would be even more difficult with HTR-type fuels of whatever composition, since the particle coatings and graphite matrix are chemically resistant and troublesome to break down mechanically.

In conclusion, if it were thought necessary to make the thorium cycle more attractive by reprocessing thorium fuel to recover and recycle U-233, this would require a significant R&D programme to work out a viable industrial process.

The first boiling water reactor experiments (BORAX-1) were carried out at the National Reactor Testing Station (NRTS) in Idaho by the Argonne National Laboratory (ANL) in 1953. This proved that in-core boiling would be stable and also verified that the void coefficient was negative and so could be used to control the reactor. The first experimental boiling water reactor (EBWR) was then built by ANL in 1956 near Chicago. Development of commercial BWRs was undertaken by General Electric (GE) and the prototype Vallecitos Boiling Water Reactor was built in 1957 near San Jose, California. This then led on to the development of commercial BWRs.

The development of the GE BWRs is illustrated in Fig. 10.10 . The first US commercial nuclear power plant Dresden 1 was a BWR1. The early BWRs had relatively low power densities and used both internal or external (steam drum) separation and were dual cycle providing direct cycle steam as well as producing steam from steam generators. The designs were simplified and pure direct cycle operation was adopted. The early BWR plants were successfully exported and adopted. Other manufacturers emerged often initially as partners of GE. BWRs were manufactured by Toshiba, Hitachi (Japan), ASEA Atom (Sweden) and KWU (Germany).

This section of Chapter 12 looks at the design of the advanced gas-cooled reactor (AGR), which was chosen by the Central Electricity Generating Board as the succession plant to the Magnox reactors in the early 1960s instead of either the pressurised water reactor or the boiling water reactor.

The commercial AGR (CAGR) was developed from a prototype built at Windscale (WAGR), which was operated by the United Kingdom Atomic Energy Authority (UKAEA). There are a number of similarities between the AGR and the Magnox designs in that both are cooled by pressurised carbon dioxide and moderated by graphite. The fuel designs are quite different, however, because the AGR operates at higher temperature and is, therefore, more thermally efficient.

The first commercial AGR commenced construction at Dungeness in Kent on the site alongside an existing Magnox station, which has since closed down. The contract for a pair of reactors was awarded in 1965 and was closely followed, around 1967, by orders for stations at Hinkley Point in Somerset and at Hunterston, a replica design, in Ayrshire. As at Dungeness, both sites already had existing Magnox reactors on them. Different designs were chosen for the next stations at Hartlepool on Teesside and at Hey sham in Lancashire in 1969 and 1970 respectively. Neither of these sites had Magnox plant. The final two AGR stations constructed were on the same Heysham site and at Torness and were ordered nearly ten years later. These are again sister stations and are of a design similar to the reactors at Hinkley Point and Hunterston. All AGRs on sites with Magnox plant were termed the ‘B’ station. The Heysham AGRs are named ‘1’ and ‘2’. All AGR stations consist of a pair of reactors in a single building. They have a common service island for fuel handling and equipment maintenance and which is either between the two reactors or at the end of the building in the case of Hartlepool and Heysham.