The operator of any nuclear plant must demonstrate that the operation of the plant is compliant with all the legislative requirements in terms of both radiological and non-radiological issues. The regulatory bodies dealing with the nuclear safety aspects of plant operation are the Nuclear Installations Inspectorate (NII) and the Environment Agency (EA) in England and Wales or the Scottish Environment Protection Agency (SEPA) in Scotland. The NII assesses the safety of the plant during normal operation with respect to exposure of the plant operators and the public to radiation, and also reviews the effect on the operators and the public of a range of fault conditions which might occur. The NII will only allow operation of the plant when it is satisfied that the plant is adequately safe and that the risks from it are acceptable.

Permission to raise power on any new plant is granted by the NII, in the form of a licence, based on the assessment presented in the final Station Safety Report.

The Environment Agencies grant an Authorisation (effectively a licence) to dispose of any gaseous, liquid and solid wastes from the site. The most significant radionuclides in any disposals are limited numerically to ensure that potential radiation doses to the public are controlled.

In LWRs (in particular PWRs), the relatively high differential pressure across the cladding wall (due to the high coolant pressure and the low pin fill pressure) in the first year or two of irradiation induces a compressive stress on the cladding, which causes the cladding to creep inwards. The resulting ‘clad creepdown’ leads to reduction of the pellet-cladding gap and, consequently, a reduction in fuel centreline temperatures. The creep is primarily irradiation induced, since the clad temperatures are not high enough for thermal creep to be significant.

14.2.4 Oxygen migration

In thermal reactor oxide fuel, the fuel temperatures are relatively low and the as-manufactured oxygen/metal molar ratio (O/M) is close to 2.00, i. e. the material is stoichiometric. Oxygen migration is then negligible. In contrast, in fast reactor oxide fuel, the fuel temperatures are relatively high and the as-manufactured O/M is usually less than two (typically in the range 1.97 to 1.99), i. e. the material is hypostoichiometric. Oxygen migration then rapidly occurs, i. e. as soon as the fuel is heated up at the start of irradiation. The oxygen migrates down the large fuel pellet radial temperature gradient, such that the local O/M increases in transitioning from the pellet centreline to the pellet rim, with the as-manufactured O/M maintained on a pellet average basis (Bailly et al., 1999). This is important because local fuel properties, including thermal conductivity, creep and fission product diffusion coefficients, are strongly dependent on local O/M — in particular, regions that have become more hypostoichiometric have a degraded conductivity, while regions that have become more stoichiometric (with an O/M closer to two) have an improved conductivity. There is some uncertainty over the dominant mechanism, or mechanisms, for the oxygen migration, which may be complex (Olander, 1976) — thermal diffusion (also known as the Soret effect) in either the solid or gaseous phase appears most likely.

16.1.1 Technical targets

Clearly, the overall aim of reprocessing is to recover uranium and plutonium efficiently from spent fuel while meeting safety targets and limiting environmental impact.

Some specific technical targets follow.

Recovery efficienciesfor uranium andplutonium

The efficiency of recovery for uranium and plutonium must be as high as possible (> 99%) in order to reduce the amount of these long-lived elements within the waste streams. As an example the efficiencies in La Hague reprocessing plant (France; AREVA) are around 99.88% for uranium and plutonium.

Specificationsfor reprocessed uranium

Reprocessed uranium is generally delivered as a concentrated solution (200-400 g/l) of uranyl nitrate for further conversion into oxides (UO2 or U3O8) or tetra — or hexafluoride depending upon subsequent use (storage, fuel manufacture or re-enrichment).

Standards are driven by re-enrichment and include norms regarding chemical and radiochemical impurities (less than 3000 ppm of volatile compounds at 850 °C other than uranium).

The specific в activity from fission products must be less than 18 500Bq/gU. The a activity, other than uranium, must be less than 250Bq/gU.

Reprocessed uranium includes all uranium isotopes from 232 to 238 except 237, which decays quickly to neptunium-237. These isotopes have a long half-life (>105 years) except for uranium-232, which has a half-life of 70 years. Through successive a and в decays, uranium-232 produces very hard and intense Y emitters.

The presence of these isotopes in smaller, though significant amounts has adverse effects in neutronic terms. Consequently, the uranium-232 and -236 isotopes must be particularly considered in calculation of enrichment levels.

In France, the first commercial reprocessing plant was for gas-cooled, graphite moderated fuel; it opened at Marcoule in 1958 (400 MT/yr). An equivalent plant at La Hague began operation in 1967 (400 t/year). The first plant for LWR oxide

fuel (UP2) started up at La Hague in 1976 (800 t/year) and a second, also at La Hague (UP3), in 1990 (800 t/year) (Fig. 16.26).

France4 chose the closed fuel cycle at the very beginning of its nuclear program, involving reprocessing used fuel so as to recover uranium and plutonium for reuse and to reduce the volume of high-level waste for disposal. Recycling allows 30% more energy to be extracted from the original uranium and leads to a great reduction in the amount of waste to be disposed of. Overall the cost of closed fuel cycle is assessed as comparable with that for direct disposal of used fuel, and conserves a resource, which may become more valuable in the future. Back-end services are carried out by AREVA NC. Used fuel storage in pools at reactor sites is relatively brief. Late in 2011, 70% of EdF’s used fuel was in used fuel pools, mostly at La Hague, 19% was in dry casks and 11% had been reprocessed.

Used fuel from the French reactors and from other countries is sent to AREVA NC’s La Hague plant in Normandy for reprocessing. This has the capacity to reprocess up to 1700 tonnes per year of used fuel in the UP2 and UP3 facilities. The treatment extracts 99.9% of the plutonium and uranium for recycling, leaving 3% of the used fuel material as high-level waste, which is vitrified and stored there for later disposal. Typical input today is 3.7% enriched used fuel from PWR and BWR reactors with burn-up to 45 GWd/t, after cooling for four years. In 2009

AREVA reprocessed 929 tonnes, most from EdF, but 79 t from SOGIN in Italy. By 2015 it aims for throughput of 1500 t/yr.

EdF has been sending some 850 tonnes for reprocessing out of about 1200 tonnes of used fuel discharged per year, though from 2010 it will send 1050 t. The rest is kept for later reprocessing to provide the plutonium required for the start-up of Generation IV reactors. Reprocessing is undertaken a few years after discharge, following some cooling. Some 8.5 tonnes of plutonium and 810 tonnes of reprocessed uranium (RepU) have been recovered each year from the 850 tonnes treated each year to 2009. The plutonium is immediately shipped to the 195 t/yr Melox plant near Marcoule for prompt fabrication into about 100 tonnes of mixed — oxide (MOX) fuel, which is used in 20 of EdF’s 900 MWe reactors. Four more are being licensed to use MOX fuel.

Used MOX fuel and used RepU fuel is stored pending reprocessing and use of the plutonium in Generation IV fast reactors. These discharges have amounted to about 140 tonnes per year, but rise to 200 tonnes from 2010. Used MOX fuel is not reprocessed at present.

EdF’s recycled uranium (RepU) is converted in Comurhex plants at Pierrelatte, either to U3 O8 for interim storage, or to UF, for re-enrichment in centrifuge facilities there or at Seversk in Russia. About 500 tU per year of French RepU as UF, is sent to JSC Siberian Chemical Combine at Seversk for re-enrichment. The enriched RepU UF, from Seversk is then turned into UO, fuel in AREVA NP’s FBFC Romans plant (capacity 150 t/yr). EdF has used it in the Cruas 900 MWe power reactors since the mid-1980s. The main RepU inventory constitutes a strategic resource, and EdF intends to increase its utilization significantly. The enrichment tails remain at Seversk, as the property of the enricher.

Considering both plutonium and uranium, EdF estimates that about 20% of its electricity is produced from recycled materials. AREVA’s estimate is 17%, from both MOX and RepU.

AREVA has the capacity to produce and market 150 t/year of MOX fuel at its Melox plant for French and foreign customers (though it is licensed for 195 t/yr). In Europe 35 reactors have been loaded with MOX fuel. Contracts for MOX fuel supply were signed in 2006 with Japanese utilities. All these fuel cycle facilities comprise a significant export industry and have been France’s major export to Japan. At the end of 2008 AREVA was reported to have about 30 t/yr in export contracts for MOX fuel, with demand very strong. However, EdF has priority.

To the end of 2009 about 27 000 tonnes of LWR fuel from France and other countries had been reprocessed at La Hague. In addition about 5000 tonnes of gas-cooled reactor natural uranium fuel was earlier reprocessed there and over 18 000 tonnes at the UP1 plant for such fuel at Marcoule, which closed in 1997.

At the end of 2008 AREVA and EdF announced a renewed agreement to reprocess and recycle EdF’s used fuel to 2040, thereby securing the future of both La Hague and Melox plants. The agreement supports AREVA’s aim to have La Hague reprocessing operating at 1500 t/yr by 2015, instead of two thirds of that in

2008. It also means that EdF will increase the amount of used fuel sent for reprocessing to 1050 t/yr from 2010, and so Melox will produce 120 t/yr MOX fuel for EdF then, up from 100 tonnes in 2009. It also means that EdF will recycle used MOX fuel.

Under current legislation, EdF is required to have made provision for its decommissioning and final waste management liabilities by 2011, but under a new bill that deadline would be deferred until 2016. At the end of 2009, EdF was reported to have EUR 11.4 billion in its dedicated back-end fund, compared with an estimated liability of EUR 16.9 billion.

France’s back-end strategy and industrial developments are to evolve progressively in line with future needs and technological developments. The existing plants at La Hague (commissioned around 1990) have been designed to operate for at least forty years, so with operational and technical improvements taking place on a continual basis they are expected to be operating until around 2040. This will be when Generation IV plants (reactors and advanced treatment facilities) should come on line. In this respect, three main R&D areas for the next decade include:

The COEX process based on co-extraction and co-precipitation of uranium and plutonium together as well as a pure uranium stream (avoiding any separation of plutonium). This is designed for Generation III recycling plants and is close to near-term industrial deployment.

Selective separation of long-lived radionuclides (with a focus on Am and Cm separation) from short-lived fission products based on the optimization of DIAMEX-SANEX processes for their recycling in Generation IV fast neutron reactors with uranium as blanket fuel. This option can also be implemented with a combination of COEX and DIAMEX-SANEX processes.

Group extraction of actinides (GANEX process) as a long-term R&D goal for a homogeneous recycling of actinides (i. e. U-Pu plus minor actinides together) in Generation IV fast neutron reactors as driver fuel.

All three processes are to be assessed as they develop, and one or more will be selected for industrial-scale development with the construction of pilot plants. In the longer term the goal is to have integral recycling of uranium, plutonium and minor actinides. In practical terms, a technology — hopefully GANEX or similar — will need to be validated for industrial deployment of Gen IV fast reactors about 2040, at which stage the present La Hague plant will be due for replacement.

Low — and intermediate-level wastes

A wide range of techniques is available for processing low — and intermediate — level waste. Because waste volume is one of the parameters that determine the price of disposal to the waste generator, volume reduction is often an important component of waste treatment. It may be achieved in a number of ways, the most straightforward of which is compaction. Here the wastes are simply forced down into a container. Super-compaction requires much greater forces because here whole waste drums are crushed into a puck-shaped form. These are usually concreted into an overpack. Another volume reduction technique is incineration. Wastes that are often treated this way are cellulosic materials and contaminated oil. The resulting ash, which will have a higher specific activity than the original waste, is usually immobilised using concrete. The final volume reduction method is re-melting. This is used for contaminated metal wastes, often encountered during decommissioning of nuclear facilities. Re-melting, which may be preceded by decontamination, aims to remove any remaining contamination in the form of slag so that the metal is clean enough to be recycled. Even if the metal is not recyclable, re-melting may still be economic because it produces a significant reduction in the volume of waste requiring disposal.

By far the commonest method of waste immobilisation is encapsulation with a cement grout. With wastes that contain magnesium or aluminium, reaction with the high pH pore fluids may generate hydrogen gas and require the container to be vented. Another technique, now not so favoured, is encapsulation in bitumen. Technologies that are gaining ground are vitrification, which can be used for a wide range of wastes,8 and pyrolysis, which has been found to be useful for organic ion exchange resins.

5.3.1 Background

Cheap, abundant and reliable energy is an essential component of modern life and governments invariably aim to have policies that will deliver secure and adequate supplies at lowest cost. One consequence of this is the regulation of electricity and gas prices and widespread subsidies on end-user fossil fuel prices — estimated at $312 billion per annum18 in 2009. Against this we have the fear of anthropogenically forced climate change, which leads governments to act to reduce CO2 emissions. Given these opposing pressures, nuclear power may seem to be a godsend: an abundant, secure and reliable source of power at reasonable cost with low carbon emissions. It is not surprising that an increasing number of IAEA member states have signalled their intention to construct NPPs in the coming decades and, notwithstanding the Fukushima accident, this seems unlikely to change.

With the notable exception of the USA, almost every NPP currently operating today was constructed under government sponsorship. This was usually in the form of direct action by a government-owned agency or utility or through the provision of government loans. Times change and the tendency nowadays is for governments to take a more hands-off approach. This greater distance does not stem from lack of interest but, rather, from the unwillingness of many governments to accept commercial risks that could be taken on by the private sector and from a belief that public sector projects are rarely models for efficiency and cost control.

The small difference in the size of the 235U and 238U nuclei results in them having slightly different ionisation potentials, slightly shifted absorption lines and forming chemical compounds with slightly different bond energies. Lasers produce beams of light at a single wavelength and therefore a very specific energy, making it possible to tune a laser so that it will preferentially interact with 235U, rather than 238U. This selective activation offers a means of differentiating between the two isotopes so that they may then be separated. Very high separation factors can be achieved, which offer the possibility of a single-stage enrichment process. Energy consumption is low, comparable with or potentially less than for centrifuge enrichment.

The technology required to manufacture and tune laser systems to the very precise wavelength necessary to interact preferentially with 235U is sophisticated but within current capabilities. Perhaps a greater challenge for deployment within a commercial facility is to ensure that the systems operate reliably and without any wavelength drift for prolonged periods, as even a fractional drift will prevent them from performing their function. The greater path length that the laser light is required to travel at large scale will also increase inefficiencies caused by factors such as absorption, reflection, refraction and diffusion as the light passes through the gas.

The activated 235U species are produced as an intimate mixture within a bulk 238U matrix, requiring that they be separated and collected efficiently if high selectivity is to be retained. This must be done quickly to avoid the activated species recombining, exchanging with bulk 2 38U isotopes or interacting with construction materials before separation can be effected. The larger the unit, the longer the residence time is likely to be and the more difficult separation is likely to become.

Laser technologies are capable of producing uranium at commercial enrichment levels in a single stage; however, a commercial plant must also have a high throughput and make efficient use of feed material. Whilst a full-scale, commercial laser plant may be more compact than an equivalent centrifuge plant, it will still be of significant size and contain a great many laser systems and optical components.

A number of different laser based enrichment technologies have been explored, notably: [11]

• Chemical Reaction Isotope Selective Laser Activation (CRISLA)

• Condensation Repression Isotope Selective Laser Activation (also CRISLA)

• Separation of Isotopes by Laser Excitation (SILEX).

In the AVLIS process, uranium metal is melted by means of an electron beam that generates a stream of uranium atoms in gaseous form. A dye laser, powered by a copper laser, is used to preferentially ionise the 235U atoms in the vapour to 235U+ as it passes an ion extractor with an applied electromagnetic field. The field draws the charged ions towards the collector where they are separated and collected as a liquid metal. Some neutral 238U atoms will coincidentally deposit as they pass the collector as will 2 38U+ ions that have become charged by exchange with 2 35U+, reducing the overall efficiency of the enrichment process. The uncharged bulk passes to a second, tails collector where it is again recovered as a liquid metal. The technology has been investigated in a number of countries, most notably the USA, where the Lawrence Livermore National Laboratory developed the process to a stage where a demonstration in 1992 produced uranium enriched to commercial levels from tonne quantities of uranium feed. USEC sought to commercialise the process later in the 1990s but abandoned it at the end of the decade as not cost effective.

The MLIS process uses UF6 in a cooled carrier gas as the feed material, which sits more comfortably within the existing nuclear fuel cycle than the uranium metal used in AVLIS. The 235UF6 is preferentially energised and then stripped of a fluorine atom to form uranium pentafluoride (UF5) using a one — or two-laser system (ultraviolet and infra-red or infra-red alone). The UF5 is not volatile and solidifies more readily within the gas stream than the UF6 so that it can be preferentially filtered from the carrier gas. The feed gas also contains a scavenger gas, such as methane, that will capture the free fluorine atoms generated during laser excitation. The technology was pursued by a number of organisations in the 1980s and 1990s but was abandoned towards the end of this period, with the notable caveat that the limited information available on the SILEX process suggests that it is related to MLIS.

I n the chemical reaction CRISLA process UF6 is mixed with a proprietary chemical reagent known as RX. A laser is used to preferentially excite 235UF6 so that its reaction rate with the RX compound is significantly increased compared to the non-excited 238UF6. The reaction product, which is enriched in 235U as a result of the increased reaction rate, may then be separated from the UF6 using techniques appropriate to the physical and chemical nature of the product. This CRISLA process was developed and patented by Dr Jeff Eerkens in the late 1970s with the technology later transferred to Isotope Technologies Inc and Cameco. Changes in market conditions led to the process being dropped in the early 1990s.

Dr Eerkens has also been involved in the development of the condensation repression CRISLA process. For this process, the feed gas of UF6 in a xenon carrier gas is cooled adiabatically down to less than 60 K through expansion from a nozzle as a supersonic jet stream. UF6 dimers are formed in the gas stream as a result. A suitably tuned laser is used to provide the 235UF6 molecules in the gas stream with enough energy to prevent them from forming dimers, thus creating a substantial difference between the mass of the non-excited, dimerised 238UF, and the laser-maintained 235UF6 monomer. This results in different radial escape rates for the two isotopes in the jet stream and allows separation using appropriately positioned skimmers.

Neither CRISLA process has been used for commercial production.

The most promising laser based enrichment process at present appears to be the SILEX process, which was originally developed by Silex Systems Ltd in Australia. Global Laser Enrichment, a subsidiary of GE Hitachi Nuclear Energy and also supported by Cameco (24% ownership), has stated that it intends to use this process as the basis for a commercial enrichment facility in Wilmington, North Carolina. The process was also the subject of a significant development programme led by USEC from around 1996 to 2002 but was not pursued to the full scale commercialisation now proposed. For reasons of both commercial and nuclear proliferation security, very little technical information on the process has been published. It is known that the feed stream is a cooled mixture of UF6 in a carrier gas with the 235UF6 preferentially excited at the 16 pm wavelength (similar to the MLIS setup). The process is based on UF6 in all process streams. Further details have not been published in the open literature at the time of writing.

The lower end plug is welded to the cladding tube, and the rod is loaded with the pellets, an insulation pellet and the spring. The upper end plug is welded on, the tube is evacuated, pressurised with helium and welded shut. The helium pressure depends on the reactor type and is 3 to 10 bar (STP) for BWR and 10 to 20 bar for PWR fuel rods. These pressures double during warming up to the operating temperature and are chosen to be high enough to provide some support against fast cladding collapse under the outside system pressure (70 bar BWR, 150 bar

PWR, but slow cladding creep-down still occurs, see Chapter 14) and to maintain a good heat conductance across the pellet-cladding gap even when the helium fill gas is diluted with low conductivity fission gas (see Chapter 14). The amount of fill gas is low enough to leave room for released fission gas without exceeding rod pressure safety limits (see Section 9.6.1).

The whole assembly and control process is highly automated. The girth weld of the end plugs is inspected by various (automated) methods, e. g., helium leak testing, computerised optical scanning of the weld surface and determination of permissible weld bulge that does not interfere with the spacer grid, X-ray fluorescence spectrography, and ultrasonic inspection.

The CANDU fuel bundle is relatively simple in structure. The fuel is designed to be compatible with on-power refuelling in a pressure-tube reactor, and like the reactor, to have high neutron economy. The bundle is small (nominally 10 cm in diameter, 50 cm long and weighing around 23 kg). These features facilitate remote handling and would, therefore, be suitable for recycling and fabrication in some advanced fuel cycles.

There are only seven components in a bundle (eight in the CANFLEX® design[21]). See Fig. 11.4 for an illustration of the components of a 37-element CANDU fuel bundle that is currently in use in many CANDU stations.

The fuel ‘meat’ is UO2 , with a density higher than in LWR fuel and with the ends of the pellets contoured (dishes and chamfers) to offset swelling and wheatsheafing (or hour-glassing, see Chapter 14).

The fuel sheath is composed of a zirconium alloy. The low burnup corresponding to the use of natural uranium fuel results in very low corrosion on the inside and outside of the sheath. The fuel sheath is thinner than in LWR fuel, which improves neutron economy and which allows the sheath to contact the fuel pellets under the pressure of the coolant and the thermal expansion of the pellets. This gives good thermal contact and reduces fuel centreline temperatures. The system is designed so that the coolant pressure always exceeds the internal fuel element pressure from gaseous fission products.

A graphite coating (CANLUB) on the inside of the fuel sheath protects against environmentally assisted cracking (EAC), sometimes called stress-corrosion cracking (SCC), which could be initiated by pellet-cladding interaction during on-power refuelling. One of the protection mechanisms of CANLUB is believed to be the gettering of corrosive fission products.

Small zironium-alloy spacer pads are brazed to the fuel sheath to prevent them from touching one another.

Larger zirconium-alloy bearing pads are brazed onto the outermost fuel sheaths to prevent them from touching the pressure tube. Both spacer pads and bearing pads increase coolant turbulence, which improves heat transfer. It is noted that in India, the appendages are welded (rather than brazed) to the fuel sheath.

Endcaps seal the element at each end.

The endcaps are welded to endplates at each end of the bundle, which hold the elements together. The bundle has to have sufficient structural rigidity so as not to fail through vibration or fretting, yet sufficient flexibility to pass through a pressure tube that has sagged as a result of age.

The CANFLEX bundle also contains additional small, non-contacting appendages brazed to the sheath to further promote coolant turbulence.

Two changes have taken place in Canada to enable more power to be obtained from the fuel bundle without exceeding power limits on the individual elements. The first was an increase in the number of fuel elements in a bundle, from 7 to 19, 28 and 37. The CANFLEX bundle, which has been qualified but is not yet in commercial use, has 43 elements arranged in rings of 1, 7, 14 and 21, with the central eight elements having a larger diameter than the outer 35. This results in a flatter rating profile across the CANFLEX bundle so that it becomes possible to generate 20% more power than the 37-element design at the same maximum linear element rating (Inch et al. , 2000). The second change was an increase in pressure tube (and fuel bundle) diameter, in going from the 19-element fuel bundle in the Douglas Point reactor (in which the pressure tube inside diameter was 82.6 cm) to the 28-element and 37-element fuel bundles in the Pickering and later reactors (in which the pressure tube inside diameter was 103 cm).

On-power refuelling increases the energy extracted from the fuel by about 25% compared to batch refuelling. This is because it improves neutron economy by avoiding the need for burnable neutron absorbers or control rods that, with batch refuelling, would be needed to suppress the excess reactivity. During on-power refuelling, a pair of refuelling machines attaches to each end of a fuel channel: new fuel bundles are inserted into one end of the channel and an equal number of old fuel bundles are discharged from the opposite end (Fig. 11.5). The CANDU 6 fuel reactor is refuelled in the direction of coolant flow, with the coolant drag pushing the fuel string down the channel. So both coolant flow and refuelling are bi-directional (coolant and refuelling direction being opposite in adjacent channels). This helps to flatten the axial power distribution, since old fuel at the end of one channel is next to new fuel in the adjacent four channels.

The refuelling rate matches the reactivity decay rate. In a CANDU 6 reactor with natural uranium fuel, about 15 bundles are replaced each day, using 8-bundle fuel shifts. Hence, two or three channels are refuelled on average each day. The number of bundles inserted at each visit of the refuelling machines to a channel is a balance between the reactivity perturbation (and resultant local power peaking) and the achievable refuelling frequency.

On-power refuelling provides the CANDU reactor with a great deal of flexibility, allowing it to accommodate different fuel types and fuel cycles. The number of bundles added during each visit of the refuelling machines can be reduced to accommodate higher enrichment fuel. It is also feasible to shuffle the bundles axially during refuelling, to shape the axial power distribution along the channel. (See Younis and Boczar (1989b) for an example of axial shuffling with LEU fuel.)

After discharge from the reactor, the used fuel bundles are stored in a water — filled bay at the station. After about six years, the decay heat from the used natural uranium fuel drops to a level which allows the fuel to be air cooled, and transferred to a dry-storage facility, if so desired. A number of such dry-storage designs are in use at CANDU power reactor sites (see for instance, several papers in CNS, 2005). The reference plan in Canada for the long-term management of used fuel is based on adaptive phased management. The ultimate endpoint of emplacement of the used fuel in a deep repository would be preceded by a long period of interim used-fuel storage, in which technical advances or changes in societal values or public policy could be accommodated (NWMO, 2005). The technical aspects of deep geological disposal have been well established. Funding is provided through a levy on the cost of electricity. The Canadian programme is managed by the Nuclear Waste Management Organization (NWMO), which includes participation from the nuclear utilities. This approach for the ultimate disposition of the used fuel has been established through extensive public consultation and includes site selection in a willing host community with appropriate geology.

77.5 Schematic showing CANDU on-power refueling (figure is copyright Atomic Energy of Canada Limited and is used with permission).

A number of intrinsic and extrinsic measures provide a high degree of proliferation resistance for all stages of CANDU reactor technology, from fuel fabrication to the handling and use of the fuel at the station, including fresh fuel receipt and storage, refuelling the reactor and used fuel management. Extrinsic measures ensure rigorous compliance with IAEA safeguards (Whitlock and Lee, 2009).

The achievable burnup in a CANDU reactor is determined by a number of factors: the fuel type (natural uranium, LEU, MOX, thorium); the core size (which determines the fraction of neutrons lost through leakage — so larger CANDU plants such as Bruce and Darlington have higher burnups than the smaller CANDU 6 reactors); and the number and reactivity worth of adjuster rods (reactivity control devices, which are discussed in Section 11.5). The burnup in a CANDU 6 reactor with natural uranium fuel is nominally 7.5 MWd/kg heavy element (HE); in the Bruce A reactor (which is larger and which has no adjuster rods) the burnup is about 8.9 MWd/kg HE.

The VHTR builds on high-temperature reactor (HTR) systems (Fig. 13.5). The system is characterized by its unique fuel form, consisting of tiny coated fuel particles embedded in a graphite matrix and located in a graphite core cooled by helium (Fig. 13.6). The refractory nature of the fuel and core materials permits core outlet temperatures higher than 900 °C (with an ultimate goal of 1000 °C). Advantages of the VHTR include its potential for high burn-up (even higher than 150-200 GWd/tHM), safety, low operation and maintenance costs, and modular construction.

The VHTR is seen as a system for the co-production of electricity and hydrogen and the supply of process heat for industrial applications (see Fig. 13.7). The

Fuel kernel

Porous buffer

Inner pyrocarbon

13.6 VHTR fuel particles and the two types of fuel elements (Brossard et al., 2009).

Power conversion unit 13.7 VHTR with electricity and hydrogen production alternatives.

high core outlet temperature allows both hydrogen and electricity to be generated with high efficiency (Romanello, 2003). The latter may use either a direct (helium gas turbine) or indirect (gas mixture turbine) Brayton cycle. Where process heat is to be supplied, however, an intermediate heat exchanger will need to be connected to the primary loop. This will also serve the hydrogen production process and could use a working fluid such as helium, a gas mixture or a molten salt. The VHTR will be developed using existing materials at first, but new more advanced materials will be needed if it is to reach its full potential.

A number of experimental and prototype HTR and VHTR units have been built, including:

• the US Peach Bottom Reactor (40 MWe, operated 1967-74)

• the Fort Saint Vrain Reactor (330 MWe, operated 1976-89)

• the Dragon Reactor (20 MWft, operated 1965-76)

• the German Arbeitsgemeinschaft Versuchsreacktor (AVR, 15 MWe, operated 1967-88)

• the Thorium Hochtemperature Reaktor (THTR, 300MWe, operated 1983-9).

Several projects to build new prototype high-temperature gas-cooled reactors are described by the Generation IV International Forum (2009). These include experimental reactors in Japan (HTTR, 30 MWft) and China (HTR-10, 10 MWth).