Gaseous swelling and fission gas release are described below. For conciseness, the term ‘fuel’ is used to refer to fuel pellets (oxide fuel) or fuel bars (metallic fuel).

Fission gas — predominantly xenon and krypton — atoms are generated uniformly within each fuel grain, each with a significant kinetic energy. Since the fission fragment range in the fuel is appreciable — of the order of 6 pm in UO2 (Noggle and Stiegler, 1960) — a significant fraction of atoms in grains close to the fuel surface are ejected into the pin free volume — this is known as recoil release, and is an athermal release mechanism. The remaining fission gas atoms come to rest in the fuel matrix, where they are effectively in solution. They are then subject to diffusion, and tend to either become trapped in intra-granular fission gas bubbles (which are nucleated in the wake of energetic fission fragments), or reach the grain boundary. The amount of fission gas that diffuses to the grain boundaries is dependent upon both the fuel temperatures and grain size, especially the former. In the case of external grains (those at the fuel surface), the fission gas atoms are released to the pin free volume. In contrast, for internal grains, the atoms that reach the grain boundary are rapidly accommodated into grain face bubbles via grain boundary diffusion. It is also possible that diffusing fission gas atoms near to the fuel surface are knocked out of the fuel by energetic fission fragments in an athermal process known as knockout release, but experimental measurements of radioactive fission gas release have shown that release due to this mechanism is negligible (Lewis, 1987). Any grain growth alters the amount of fission gas that reaches grain boundaries via two effects: the increase in the average distance over which fission gas atoms have to diffuse to reach a boundary; and the accumulation of intra-granular gas at moving grain boundaries (a process known as grain boundary sweeping). The first effect tends to decrease the rate of fission gas release to the pin free volume, while the second effect tends to increase it.

The intra-granular and grain face bubbles (which are actually quasi-crystallites, and not bubbles in the traditional sense) grow as fission gas atoms and vacancies diffuse to them and as intra-granular gas is swept up by grain boundary migration. However, both gas atoms and vacancies (which together comprise the quasi­crystallite material) are also ejected from the bubbles by energetic fission fragments in a process known as irradiation-induced re-solution. Thus, the size of the bubbles is governed by the net result of these two competing processes. Because of the large difference in the volume occupied by a gas atom in the fuel matrix and in a bubble, the presence of the intra-granular and grain face bubbles leads to gaseous swelling of the fuel.

In steady-state conditions, the radius of the (spherical) intra-granular bubbles tends towards an equilibrium value, while the grain face bubbles tend to grow inexorably. The inter-granular bubbles are initially lenticular, but as they grow and coalesce they can, if the imposed stress is small enough, become first elongated and then vermicular (‘worm-like’) (Barker et at., 2009). The grain face bubbles that intersect a grain edge are subsumed into grain edge bubbles. Once the areal density of fission gas on the grain face reaches a critical value, the isolated grain edge bubbles become long enough that they interlink to form tunnels to the fuel surface. The gas in these interlinked tunnels is then rapidly vented to the pin free volume. Once vented, the tunnels may close due to sintering, but they can re-open with time as more fission gas arrives at the grain boundaries. The dependence of bubble morphology and size on temperature, fission rate, burnup, etc., means that gaseous swelling strains can vary by a large amount depending on the precise irradiation conditions.

To summarise the above: gaseous swelling occurs due to nucleation and growth of both intra — and inter-granular fission gas bubbles; and fission gas release (release of fission gas from the fuel pellets or bars to the pin free volume) occurs via (i) recoil (athermal), (ii) diffusion to free surfaces (both athermal and thermal) and (iii) diffusion plus interlinkage (both thermal and athermal) mechanisms. A fourth (thermal) release mechanism — bubble migration — occurs at high temperature (above ~ 1800°C (Turnbull, 1976) for UO2 fuel). Finally, there is the possibility of enhanced fission gas release from thermal reactor oxide fuel at high burnup due to saturation of the fuel matrix with fission gas, either locally at the pellet rim (see the discussion in 14.2.14 on formation of high burnup structure), or throughout the fuel pellets (Sontheimer and Landskron, 2000).

The fission gas released to the pin free volume has a significantly lower thermal conductivity than the helium fill gas. The gap conductance is therefore noticeably reduced, leading to an increase in fuel temperatures. The released fission gas also increases the pin internal pressure. If the fission gas release becomes overly large, the high pin internal pressure can cause creepout of the cladding. If the rate of creepout is greater than the rate of fuel swelling, there is an opening of the gap between the pellets/bar and cladding. This increases fuel temperatures, which in turn can lead to further fission gas release. This positive feedback mechanism has the potential to cause rupture of the cladding.

The fission gas fractions releasedby each of the four mechanisms, and therefore their relative importance, are dependent upon fuel type. All mechanisms are important for fast reactor, HTR and CANDU fuel, where fuel temperatures are high. In contrast, only the first mechanism is important in Magnox fuel, where fuel temperatures are low. Finally, only the first three mechanisms are important for LWR and AGR fuel (diffusion plus interlinkage is dominant once interlinkage occurs), where fuel temperatures are intermediate. The magnitude of fission gas release is dependent upon fuel type, fuel pin design, the irradiation history (in particular the fuel burnup), and the fuel grain size. Typical end-of-life values for LWR (Johnson et al. , 2004), AGR (Barrable et al., 1997) and CANDU (Floyd et al., 1992) UO, fuel are in the range 0.1 to 10%, whereas typical end-of-life values for metallic and oxide fast reactor fuel are 60% (IAEA, 2003b) and 80% (Maeda et al., 2005), respectively. Both gaseous swelling and fission gas release are significantly enhanced during any transient increases in pin power.

Civil reprocessing of spent fuel utilizing the PUREX process has been successfully practised on a commercial scale for over 40 years without any occurrences of the diversion of nuclear materials.11 These operations have been for spent fuel management and for the recovery of the uranium and plutonium for recycling as UOX and MOX fuel for light water and fast reactors. Such a combination of spent fuel reprocessing and recycling may lead to benefits in ultimate waste disposal primarily due to reduced volumes of reprocessing HLW (the volume reduction factor of spent fuel to reprocessing HLW is a 4:1).

Irradiated nuclear fuel from research reactors was first reprocessed in the 1940s using pyrochemical and precipitation processes. These separation methods were soon replaced by the solvent extraction process (hydrometallurgy), which is better suited for continuous, large scale, remote operation, and can separate the three main streams of radionuclides (uranium, plutonium and waste, i. e. fission products and actinides). Different solvent extraction systems were explored before the discovery of an efficient extraction system. The combination, known today generically as PUREX, soon replaced all earlier solvent extraction methods because of its high performance in industrial-scale plants. PUREX utilizes the

15.15 Disposal facility concept for the CANDU heavy water reactor type of fuel (Courtesy of Nuclear Waste Management Organization in Canada).

extractant tributyl phosphate (TBP), mixed in a largely inert hydrocarbon solvent. The first plant for reprocessing based on the PUREX technology was built in Belgium in the 1960s. In the 1970s, there was some expansion of reprocessing capacity and its application to fuels from various types of reactor. In the 1980s, due to proliferation and other concerns, the strategy moved to a once-through cycle with the disposal of spent fuel. However, several countries including France,

Japan, UK, Russia and India continued to further develop, improve and adapt the PUREX technology. In France it was used for MOX fuel fabrication, in Russia for U recycling for the RBMK reactors and in India for U recycling of PHWR fuel and MOX to make fast breeder reactor fuel (FBR). Further consideration of spent fuel reprocessing has to be done within the current circumstances where long­term storage seems to be an interim strategy that will have to be combined with future advanced reactors and nuclear fuel cycles.

Sustainability is a major driver in developing advanced nuclear fuel cycle technologies.

Developments in advanced reprocessing technologies are directed toward the following goals:

• Reduction of reprocessing cost in comparison to the current PUREX process costs and in comparison to direct disposal costs of the once-through fuel cycle.

• Recovery of all actinides and long-lived fission products to reduce the volumes and radiotoxicity of the radioactive waste for disposal and hence a decrease in the expense of waste disposal and an increase in the long-term safety of the repository.

• Creation of flexible technologies that are adaptable for changing conditions and requirements such as new designs and materials for fuel and reactors of the third and fourth generations.

• Through a well understood reduction of safety risks and proliferation risks, to render nuclear power more acceptable to the public.

There are several national and international initiatives supporting the development of advanced nuclear fuel cycles. The International Atomic Energy Agency (IAEA) started the INPRO initiative and multinational approach in nuclear fuel cycles. There is also the generation IV international forum (GIF), the Russian initiative on development of international nuclear centres, the USA initiative Global Nuclear Energy Partnership (GNEP, lately renamed the IFNEC-International Framework for Nuclear Energy Cooperation) and some others.

Currently available and developing reprocessing technologies can be divided into groups according to their stage of maturity: [30]

• Evolutionary technologies (generation III reprocessing facilities) based on aqueous separation methods have been successfully tested and are ready for industrial implementation. The objective of these technologies is the co-management of U and Pu (or U-Pu-Np). One of the key features of these processes is that separation of Pu does not take place, which significantly reduces proliferation concerns. Furthermore, an integrated facility for reprocessing and fresh fuel re-fabrication can be applied. There is also flexibility in the processing systems to allow fabrication of MOX fuels both for LWRs and FBRs. There are also advantages like enhanced MOX fuel performance due to high homogeneity of fabricated fuel and the possibility of selective separation of some minor actinides and fission products. Figure 15.16 shows as an example the block diagram of the COEX reprocessing system developed in France.11

• Aqueous processes using new extractant molecules will provide two possible options for separation of actinides. One possibility is selective separation of minor actinides (MA) for interim storage allowing the postponed decision on transmutation in heterogeneous mode, either in fast reactor blankets or in accelerator driven systems (DIAMEX-SANEX is under development in France, TALSPEAK in the USA, TOGDA in Japan)

• The other option is group actinide separation using an integrated fuel cycle (on-line fuel reprocessing and fabrication) with the prospect of their homogeneous recycling in fast reactors (GANEX in France, UREX+ in the USA, NEXT in Japan).

• Innovative methods based on pyrochemistry (‘dry methods’) will allow reprocessing of different types of highly radioactive fuels such as metals, carbides, oxides or nitrides with high content of fissile material (in the fabrication of dedicated fuels for transmutation purposes or minor actinides targets), or fuels with high burn up. The advantage of the pyrochemistry method is that it is performed in inorganic media. The result of this is that the process is less sensitive to radiation effects allowing early reprocessing of fuel after discharge from a reactor. Furthermore, there is a low criticality risk compared to aqueous methods where water is an efficient neutron thermalization media. These ‘dry’ methods would also be very suitable for reprocessing as designed for molten salt reactors.

• A combination of hydro — and pyro-processes may have some advantages from each of the two processes but their efficiency may be affected by discontinuity between the two steps in the process.

• Other innovative processes are still more or less at the laboratory scale and include a process using Freon fluid or supercritical CO2 extraction, processes based on chromatographic methods and processes using precipitation methods.

In conclusion, there are a number of options for recycling of spent nuclear fuel.11 Some, including those that avoid separation of the pure plutonium stream, are at an advanced level of maturity. These could be deployed in the next generation of industrial scale reprocessing plants, while others (such as ‘dry’ methods) are at pilot scale, laboratory scale or a conceptual stage of development. Measures to improve the environmental protection of commercial reprocessing plants over the past 20-30 years have greatly reduced emissions and waste volumes11 in currently operating reprocessing plants.

The deployment of multi-national fuel cycle centres, operating under an international framework and most effectively implemented in those countries with a sufficiently large civil nuclear energy infrastructure, can serve to ensure a sustained supply of nuclear fuel and related services under conditions in which the risk of proliferation of technologies related to the production of nuclear weapons is minimized. Reprocessing of spent fuel will be an important function of these centres.

The next generation of spent fuel reprocessing plants will likely be based on aqueous extraction processes. The physical design of these plants will have to incorporate effective means of material accountancy, safeguards and physical protection. Innovative reprocessing technologies must be developed for the reprocessing of fuel types that will be used in future and that may be substantially different from the UO2 and MOX ceramic type fuel used today.

16.9.1 United Kingdom

In the United Kingdom, 1 the Windscale (later Sellafield) plant for reprocessing of Magnox fuel began in 1964 (1500t/year) and is expected to close when all Magnox fuel has been reprocessed around 2016. The thermal oxide reprocessing plant (THORP) accepts oxide fuel from gas-cooled and light-water reactors; it began operations in 1994 (900 t/year). As of early 2010, it had treated about 6000 tonnes of used fuel for overseas and domestic customers. Of this, 2300 tonnes was domestic used AGR fuel. A further 6600 tonnes arising to the end of the AGR operating lifetimes will need to be treated or stored, depending on the outcome of a review of used oxide fuel management strategy. Less than 700 tonnes of fuel from overseas customers remains to be reprocessed. It appears likely that Thorp will operate to 2020, according to the NDA’s revised strategy due to be finalized early in 2012 (Fig. 16.25).

Reprocessing activities at Sellafield are now undertaken by Sellafield Ltd on behalf of International Nuclear Services, which is owned by the NDA (Nuclear Decommissioning Authority).

Practical experience has shown that the efficiency of radioactive waste management is greatly improved if the system of waste classification is aligned with the proposed method of disposal. Because near-surface disposal is reserved for short­lived waste (half-life less than about 30 years), this requires that the system should reflect the half-life of the radionuclides in the waste. The classification scheme recommended by the IAEA5 follows this model (illustrated schematically in Fig. 18.1) with six specific categories, as briefly described in Table 18.1.

Table 18.1 Brief description of the various waste categories in the IAEA scheme

Category Description/typical Disposal route

maximum specific activity

Can be treated as non­radioactive

A key feature of the IAEA classification scheme is that it is based on the intended disposal route. For this purpose, short-lived wastes are defined as having half-lives of ‘less than about 30 years’. This allows caesium-137 (half-life 30.17 y) to be classified as short lived.

Fluctuations in demand and the possibility of unscheduled outages make it essential to have plant in reserve. This is always the case regardless of what kind of generators contribute to the overall capacity. Where a system includes intermittent sources, however, the required level of reserve capacity will be greater. This additional capacity is not specifically dedicated to covering for loss of intermittent supplies so that an investor in wind power, say, would not need to concern himself with this. Nevertheless, it does represent an additional cost on the system and is, therefore, a hidden subsidy for intermittent suppliers.

For wind turbines, back up is most obviously needed for those times when peak demand coincides with very low wind conditions. When the percentage of total capacity provided by wind (so-called wind penetration) is small, any shortfall can be readily made up by boosting the output of thermal generators or increasing imports. As penetration increases however, it becomes progressively more difficult to bridge the gap. Recent modelling indicates that at 20% wind penetration, the UK would need backup equivalent to almost 50% of the firm wind capacity defined as the installed (i. e. nameplate) capacity multiplied by the average availability.17

Gas centrifuge technology again makes use of the mass difference between 235UF6 and “8UF6 to promote separation. The process entails UFi gas being fed into a centrifuge unit spinning at very high speed, with the wall of the centrifuge acting as the rotor. The rotation of the gas applies an acceleration to the gas molecules in the direction of the centrifuge wall, with the greater force exerted on the molecule with the higher mass so that the more massive molecules concentrate at the centrifuge walls, whilst the less massive molecules concentrate closer to the central axis of the unit. The partially separated gas is then encouraged to circulate along the centrifuge axis using a thermal gradient. ‘Scoops’ are used to draw off an enriched product stream and a depleted tails stream. A diagram of a single centrifuge unit is shown in Fig. 7.5 . The centrifuge sits above the motor which drives the rotor. There is also a magnetic bearing at the top of the centrifuge through which the inlet and outlet pipes pass. The whole centrifuge is housed in a casing, with the space evacuated to reduce friction to a minimum.

The separation factor achieved by a single centrifuge unit is much greater than for a single diffuser unit, in the range 1.2 to 2.0 depending on the sophistication of the centrifuge. The enrichment required for nuclear fuel can still not be achieved with a single machine, however, and the feed rate is quite low so that a cascade containing many machines linked in series and parallel (see Fig. 7.4) is needed to achieve the desired output and to reduce the 235U level in the tails to economically viable levels. A commercial scale gas centrifuge plant is likely to have tens of thousands of separate units in a number of cascades. The greater separation factor that can be achieved by a single centrifuge still means that the footprint is likely to be smaller than for a gaseous diffusion plant with the same SWU rating. Crucially, the power consumption of a centrifuge plant, although still significant, is in the region of 2% of that required for gaseous diffusion.

The performance of a centrifuge increases with its height (axial length) and rotational speed. Construction materials are therefore required that are both strong and resistant to UF. . Older models used aluminium alloys or maraging steel, whilst later models use carbon fibre. Centrifuge units spin at speeds that can

approach the speed of sound, putting considerable stress on the machine, particularly the rotors and bearings. These parts of the centrifuge require advanced design and manufacturing capability, as does the electrical drive system, which must provide very precise speed control. The vast majority of the world’s centrifuge enrichment capacity is based on either Russian or ETC designed machines. The American Centrifuge being developed by USEC is designed to have a higher output than existing machines, although only time will prove whether it has comparable reliability.

The basic philosophy for running a centrifuge is to vacuum it down, start it spinning, feed it with UF6 and leave it. The units are kept at constant temperature and protected from impact and other physical interference and allowed to run 24 hours a day, 365 days a year without maintenance or interference. Any adjustment to the performance of a cascade in terms of product and tails enrichment is usually made by altering the feed and bleed between machines within the cascade, rather than changing the way that the machines operate. Should a machine fail, for example if a drive motor fails, then operation of the cascade is adjusted to accommodate it rather than attempting repair.

Modern centrifuges are an engineering marvel. They operate continuously and without maintenance yet rarely fail, despite the extreme stresses they endure. ETC machines on UEC sites run with a failure rate of less than 1% per year and many centrifuges have been operating continuously since they were first brought on line, some as long ago as the early 1980s.

The fuel rod components are predominantly made of zirconium alloys. Zirconium is a ductile metal with mechanical properties similar to those of titanium and austenitic stainless steel. Its alloys combine very low neutron absorption with good corrosion resistance in power reactor conditions. Stainless steel was tried for the cladding tube at the beginning of the nuclear era, but was abandoned because of stress corrosion cracking problems and high neutron absorption compared to zirconium.

Zirconium is produced from zirconium ore or sand (zircon, ZrSiO2 ) in two principal ways, namely by chemical reduction of the oxide (sponge zirconium) and by electrolysis of zirconium halides in a salt melt (Neikov et al., 2009). Although the two methods produce rather pure zirconium, the respective products differ in their content of impurities. It has been found that this can lead to differences in the behaviour of the alloys produced from them. Yegorova et al. (2005) studied E110 (the Russian cladding tube alloy traditionally manufactured from electrolytic process zirconium) in loss-of-coolant temperature conditions. They found that material produced from the alternative sponge zirconium showed a significantly reduced oxidation rate and tolerated more oxidation (equivalent cladding reacted, ECR) before the material completely lost its ductility.

Reactor grade zirconium (ASTM B349, 2009) suitable for use in nuclear applications is characterised by its low neutron absorption cross section achieved by removal of hafnium, which occurs in the same mineral in quantities of 1.5 to 4% (the hafnium is used for control rods). The impurities remaining in nuclear grade zirconium after the hafnium extraction process are given in Table 9.3 (Moulin et al., 1984).

For in-core applications, zirconium cannot be used as a pure metal, but has to be alloyed in order to obtain better corrosion resistance. The nuclear era started with alloys called Zircaloy-2 (Zry-2) for BWRs and Zircaloy-4 (Zry-4) for PWRs.

Table 9.2 Typical dimensions of cladding and fuel pellets Feature PWR BWR 14 x 14 15 x 15 16 x 16 17 x 17 18 x 18 9×9 10 x 10 Cladding outer diameter (mm) 10.16-11.18 10.75 9.14-10.75 9.50 9.50 11.00-11.20 9.84-10.28 Wall thickness (mm) 0.570-0.725 0.620-0.725 0.570-0.725 0.570 0.64 0.70-0.71 0.605-0.620 Pellet — cladding diam. gap (mm) 0.170-0.210 0.170-0.190 0.160-0.190 0.156-0.170 0.170 0.180-0.200 0.150-0.170 Pellet height (mm) 10.0-12.6 10.0-12.6 9.8-11.0 9.8-13.5 9.0-9.8 10.0 10.0-10.5

The composition is Zr/1.5% Sn/0.15% Fe/0.1% Cr/0.05% Ni for Zry-2 and Zr/1.5% Sn/0.2% Fe/0.1% Cr for Zry-4 (Schemel, 1977). Later, increasing demands on the fuel cladding due to extended burn-up and in-core residence time led to the development of improved alloys especially for use in PWRs. A variety of products are available from different vendors, e. g. ZIRLO (Westinghouse), M5 (AREVA) and MDA (Mitsubishi). They are characterised by containing 0.5-1.0% niobium, which is not present in the Zircaloys. The Russian E110, which is used in VVER and RBMK reactors, contained 1% Nb from the very beginning. Other vendors developed claddings with an outer layer of corrosion resistant material on a Zircaloy substrate as explained in the next section.

The main emphasis in the development of ALWRs has been on the production of large plants. This provides benefits in terms of economies of scale as well as minimising the number of sites required. However, there has been a renewal of interest in small and medium reactors (SMRs); SMRs are generally defined as being reactors of less than about 600 MW(e).

The renewed interest in SMRs is mainly focussed on issues associated with the introduction of nuclear generation into countries or areas with smaller distribution grids, and the lower initial investment costs. The smaller SMRs can effectively be factory constructed. One example is barge-mounted units, which can be constructed in a shipyard and then towed to their final location. SMRs also tend to be simpler than large reactors and can make greater use of passive systems.

As with all electricity generation plants the economics of SMRs is important. On one hand the small size mitigates against economies of scale while the simpler designs and the ability to factory construct modular plants should reduce costs. One key consideration could be the initial investment cost. The smaller unit price may make it easier to provide the investment needed to get a programme of reactors started as well as making them more appropriate for markets where the rate of growth of demand is lower.

10.10 Sources of further information

The best sources of further information on LWRs are the websites of the international organisations involved in nuclear power because the information is regularly updated. In particular the IAEA (http://www. iaea. org/) provides safety standards and design requirements that form the basis for the national regulation of most countries. In addition they also provide extensive information on nuclear technology for both existing and developing reactor designs. OECD’s Nuclear Energy Agency (http://www. oecd-nea. org/) also provides extensive information, as does the World Nuclear Association (http://www. world-nuclear. org/).

Another excellent source of information on LWRs is the USNRC (http://www. nrc. gov/). This includes training material as well as information relating to existing and proposed plants. For the new plants this is not confined to US plants since it contains information on all the plants that are going through the design certification process for new plants. Similarly there is considerable information available on the plants going through the UK Generic Design Assessment process via the Office of Nuclear Regulations Website (http://www. hse. gov. uk/ newreactors/index. htm).

GIF has proposed an integral fuel cycle, with the reprocessed spent fuel from LWRs forming part of the feed for fast reactors (FRs). Spent fuel from FRs would then be reprocessed in situ (i. e. inside the same installation although outside the reactor). All HMs would be recovered together (i. e. without chemical separation of the different elements) and reused to produce new fuel for the same FRs (this is known as a multiple homogeneous recycle system), while fission products (FPs) would constitute the final waste (Generation IV International Forum, 2002).

There are, however, a number of challenges. There is a need to develop a cost — effective method to treat highly radioactive materials and to achieve efficient extraction of HMs (at least 99.9%). There are also radioprotection issues in treating significant quantities of MAs (particularly Cm, due to its strong у and neutron emissions) alongside significant quantities of other HMs. One solution is to recycle U, Pu and Np and, if appropriate, Am rather than the whole HM group. Np can be partitioned via the PUREX process, although this procedure has not yet been developed on an industrial scale. Instead, it may be best to store Cm until it decays into Pu, because 244Cm has a half-life of only 18 years or so.

Separating Cm from Am also poses challenges because the two elements exhibit similar chemical behaviour, making it potentially simpler to store Cm and Am together. Recovering Am and Cm would perhaps be viable in smaller dedicated facilities, where they could be heterogeneously recycled for critical reactors or for accelerator driven systems (ADSs). It is important to note that Cm recycling is very difficult to manage because it involves the creation of non­negligible amounts of 252Cf, an extremely strong neutron emitter (far more so than Cm itself, see Table 13.5) (Bomboni et al., 2009b). More research is also needed to investigate the possibility of recycling Am without Cm. Separating Cm and Am is a difficult procedure, and might not be particularly effective in terms of reducing radiotoxicity. Am reprocessing only reduces the long-term radiotoxicity by a factor of 10 or less (compared to a route without Am reprocessing), because Cm is produced by neutron capture. Finally, building the dedicated facilities needed for Am and Cm recycling might prove uneconomical.

A single reactor is unlikely to be sufficient for the burning of all the HMs. Successful transmutation is more likely to be achieved by a chain of reactors, each

performing different tasks. LWRs would be the first link in any possible chain, because their reliability has been proven internationally and LWR SNF is rich in fissionable elements. Nevertheless, for technological and neutronic reasons it is impossible to burn HMs completely in LWRs. Instead, FRs can substantially increase the availability of nuclear fuel through exploiting Pu by breeding 238U. The fast spectrum allows transmutation of both even-numbered Pu isotopes and MAs, due to its good neutron economy. The use of new TRU-based fuels will need careful investigation, focusing particularly on the dynamic behaviour of the core. Introducing large fractions of Pu and MAs tends to worsen safety parameters such as the fuel temperature coefficient (FTC) and the effective delayed neutron fraction eeff). Cores will need to be designed with neutron economy in mind, and should be able to reach and maintain criticality with small fractions of fissile Pu.

Given a core power distribution and fluid boundary conditions at the core inlet and/or outlet, a core thermal-hydraulics code predicts the three-dimensional distributions of coolant velocity, coolant pressure and coolant enthalpy within the core. The distributions of other subsidiary quantities such as coolant temperature, pin surface temperature, coolant density, mass velocity, thermodynamic quality, void fraction (for LWRs), departure from nucleate boiling ratio (for PWRs) and critical power ratio (for BWRs) may also be calculated. The predictions can be either for a single statepoint (i. e. for a single instant in time) or for multiple statepoints (i. e. for an evolution with time of the core). Steady-state codes assume equilibrium conditions for each statepoint, whereas transient codes accurately model the time variations in the fluid conditions (given appropriate time-varying core power distributions and fluid boundary conditions).

For simple analyses, the core power distributions and fluid boundary conditions are typically obtained from technical specifications. For more detailed analyses, the core power distributions and fluid boundary conditions are generally taken from the output of a whole core neutronics code and a system thermal-hydraulics code, respectively.

The code predictions are generated by solving the mass, momentum and energy conservation equations that govern the fluid flow. This is typically achieved by discretising the core into a number of notional ‘subchannels’, i. e. interconnected parallel flow channels, which extend across the full length of the core (in which case the code is often known as a ‘subchannel code’). The subchannels are further discretised in the axial direction into subchannel nodes. The mass, momentum and energy transfer between nodes can then be evaluated in a manageable way by the solution of the appropriate matrix equations (with various approximations). In the hot regions of the core, where the thermal-hydraulic conditions are of most interest, subchannel boundaries are typically delineated by the loci between the centres of adjacent pins. In the colder regions, where conditions are less important, subchannels can be larger, with boundaries encompassing multiple fuel pins and/ or assemblies.

With respect to water-cooled reactors, older core thermal-hydraulics codes tend to solve the mass, momentum and energy conservation equations for a homogeneous fluid, with empirical models to determine void fraction as a function of fluid quality (the fraction of the total mass flow rate that is vapour in a vapour-liquid two-phase flow) and to calculate liquid-vapour slip (relative motion of the two phases). In contrast, newer codes tend to solve the two-fluid equations (i. e. two sets of equations, one for each of the liquid and vapour phases), which is a more accurate, but also a more time-consuming, approach, which negates the need for the empirical models just described. However, the newer codes still require some empirical models, in particular, to take account of subcooled boiling.