Largely because of political concerns over nuclear proliferation in some countries, there has been some interest in cutting the total inventory of plutonium: first by reducing the amount of plutonium that is produced in nuclear fuel and, secondly, by utilizing the existing stockpile of separated plutonium, including ‘military’ plutonium. In the first case, thorium offers useful possibilities because it can be used as fertile material without producing plutonium; of course, it produces instead another fissile material, U-233, which can also raise proliferation concerns but mitigating measures are available, as described in Section 8.4.1. In fact, if thorium is used with medium enriched uranium (< 20%), the reactor will still produce plutonium but in lesser amounts than the standard uranium-plutonium cycle. If it is used with HEU (which would itself raise proliferation concerns), plutonium production would be very low. But if only pure U-233 is used as a fissile material in combination with thorium, then plutonium production is essentially zero. Thus, extensive use of a thorium cycle would provide a significant reduction in the rate of plutonium accumulation.

With respect to the utilization of existing plutonium stockpiles, the standard solution is the manufacture of MOX fuel for LWRs, many of which can be operated with 100% MOX. The same may also be true for HTRs. Of course, while these cycles will reduce the stockpile of separated plutonium, they will also produce fresh, if unseparated, plutonium. Thorium-plutonium fuels, on the other hand, could be deployed in various reactor types (see Section 8.2.3) and would very effectively reduce the stockpile without, in an ideal Th/Pu-fuel option, creating fresh plutonium. In the case of ‘military’ plutonium (i. e. weapon-grade plutonium now being declared excess to military needs in the United States and Russia), it has been proposed by the US that HTRs would allow plutonium to be burned in a very efficient manner. This solution is now under study in Russia, in cooperation with several countries. Thorium matrix fuel is a possibility and, although it is not strictly required in this case, the elimination of uranium from the HTR fuel would prevent the formation of fresh plutonium.

The majority of PWR designs use unshrouded fuel assemblies. In these the core effectively consists of an open array of fuel rods. The fuel assemblies consist of square arrays of pins. The number of pins and the pin diameters vary but a typical 17 x 17 fuel assembly is illustrated in Fig. 10.9. The fuel rods are held on 17 x 17 grids. This provides 289 positions. The central position is reserved for in-core instrumentation and another 24 positions are occupied by thimble tubes. These provide part of the skeleton of the fuel assembly and are structural elements, which are joined to the top and bottom nozzles as well as the spacer grids. The reactor control rod cluster assemblies (see Fig. 10.9) are inserted into these tubes in some of the assemblies. The control rod positions are fixed by the design of the plant but as all the fuel assemblies can accommodate RCCAs this allows flexibility in the placement of the assemblies in the core loading pattern.

The fuel rods themselves consist of zirconium alloy tubes into which are inserted pellets of enriched UO2. The rods incorporate a gas plenum to accommodate fission gases produced during irradiation. In general two lengths of fuel rod are commonly used, corresponding to active core heights of 12 foot (‘standard’) or 14 foot (‘XL’), or their metric equivalent. In the Westinghouse design these cores can be accommodated in the same vessel. The XL core needs a longer core barrel, which extends into the lower head region. PWRs can also use mixed oxide (MOX) fuel where plutonium oxide provides the initial fissile loading rather than 235U. Because plutonium has a lower delayed neutron fraction than uranium the shutdown margin is affected and if more than about 30% of the core is loaded with MOX then additional RCCAs are required.

The enrichment used depends on the target burnup and the fuel cycle length. Originally PWRs were designed to run on a 12 month cycle and achieve a burnup of about 30 GWd/t. Each fuel assembly stayed in the core for three cycles and a third of the core would be changed at each refuelling outage. The factor determining the burnup and hence dwell time was largely the performance of the fuel cladding. For all plants which refuel periodically rather than on-load, the reactor must be designed to cope with a number of failed fuel rods, since it would be uneconomic to shutdown, to replace fuel failures, on an individual basis. At one time the design basis was specified in terms of being able to operate with a very small percentage (~0.25%) of fuel failures. In practice it is now controlled by limiting the maximum allowable activity levels in the primary coolant and operators demand very high fuel reliabilities. Operating with failed fuel makes maintenance more complex and increases operator radiation exposure. Cladding materials have been developed to give greater resistance to radiation effects, which has allowed burnups to be extended to more than 60 GWd/t. Modern cladding materials are based on zirconium alloys. Stainless steel has been used in the past, in some fuel, but the performance of zirconium seems to be superior.

Within the constraints provided by the fuel design, the actual fuel cycle used is largely a matter of economics. Long fuel cycles mean that the ratio of generation time to refuelling time is increased. However, increasing the cycle length increases the enrichment required and also increases the number of fuel assemblies which have to be replaced at each refuelling. Small batch sizes tend to allow more efficient fuel utilisation. The replacement of individual assemblies when they reach optimum burnup is ideal, but can only be practically achieved by on-load refuelling, which is not possible for PWRs. Thus the fuel costs tend to be higher for long fuel cycles, but the average outage costs are lower. In addition there is a practical limit on the maximum enrichment currently used. Most fuel fabrication plants and transport containers are designed for enrichments of up to 5%.

Cycle lengths of between 6 months and 2 years are currently used. The short cycles are used by some German Konvoi plants, which were designed to allow rapid refuelling. The commonest cycle lengths are 12 and 18 months. Eighteen months is widely used, particularly in countries which have peaks of demand in both summer and winter, since the outages can be alternated between spring and autumn when demand is lower.

Safety case issues have much in common with those of the AGR, discussed further in Section 12.4. Issues specific to Magnox plant include:

• air ingress (leading to oxidation of the fuel and core) and loss of cooling capacity following failure of boiler ducts (not possible where boilers are contained within the pressure vessel)

• brittle fracture of the pressure vessel

• single — or multi-channel fires involving the oxidation of cladding and fuel (and potentially channel blockage)

• reactivity faults, including local removal of control rods

In general, gas-cooled thermal reactor designs are much less sensitive to loss of coolant sequences than are water-cooled reactors due to the combination of the inherently low heat transfer between fuel and coolant combined with the very high thermal capacity of the core including moderator. This makes rapid response much less of an issue than in PWR and BWR designs.

Fission products, which include solid, volatile and gaseous species, accumulate in the fuel as irradiation proceeds. Since each heavy metal atom that fissions is generally replaced by two fission product atoms, the accumulation of fission products causes the pellets/bars to swell. The fission products also degrade the thermal conductivity of the fuel. However, the behaviour is complicated by the fission product chemistry, the different ways in which the fuel matrix accommodates the fission product atoms (occupying lattice positions, as interstitials, as separate phases, etc.), and the migration of the volatile and gaseous species. Swelling is a particular problem in metallic fast reactor fuel (IAEA, 2003b), in Magnox fuel (Harris and Duckworth, 1982) and in carbide and nitride fast reactor fuel (Bailly et at., 1999). The overall swelling is often divided into two components: ‘inexorable swelling’ due to solid fission products and volatile fission products that are solid at the fuel temperatures of interest; and ‘gaseous swelling’ due to gaseous fission products and volatile fission products that are gaseous at the fuel temperatures of interest. Gaseous swelling is described in more detail below.

At the end of 2010 there was still no spent fuel repository in operation for power reactor fuel. The only similar project and operational facility that can provide lessons for spent fuel management facility is the waste isolation pilot plant (WIPP) in the US. This is a repository for transuranic waste of military origin operating in New Mexico.10

The main characteristics that influence spent fuel and high-level waste (HLW) disposal are the content of the long-lived radionuclides in the spent fuel and HLW, their heat generation and radiation levels. Heat generation limits the amount of waste that can be disposed in a given volume of rock. High radiation levels require that all radiation waste handling is shielded and uses remote handling systems. The amount of long-lived radionuclides requires the safety of the repository to be considered for tens of thousands of years. The main design objective of any radioactive material repository is to confine waste and isolate it from the environment. Furthermore, the fundamental task of any repository design, operation and performance is to demonstrate compliance with all applicable laws and regulations. Many of these are developed under the principles and guidelines developed by, or under, the auspices of the International Atomic Energy Agency (IAEA). Disposal facilities for spent nuclear fuel and HLW are typically designed for a deep geological formation with adequate rock characteristics. The reason for this is that adequate long-term safety needs to be provided without reliance on active controls and continuing maintenance of the closed facility. Geological repositories are therefore designed to be passively safe and with the idea that no indefinite institutional control is required to ensure safety. During the operating period of the disposal facility, institutional control will have to be maintained to provide assurances on safety, safeguards and security. One of the key problems with the disposal of nuclear fuel is the long-term radio toxicity of its radionuclides, in particular actinides. Figure 15.13 shows the radiotoxicity of radionuclides in spent fuel over the long time periods relevant to spent fuel disposal. The isolation of nuclear fuel to prevent release into the environment is essential.

The other issue for spent fuel disposal that has to be taken into account in choosing the location and geology of the disposal site and the design of engineered barriers is the decay heat over long periods of time. Figure 15.14 shows decay heat and its contributors over the long periods of time relevant to disposal. The third aspect of spent fuel disposal safety that is to be taken into account in the

design of engineered barriers is contact with water and demonstrating that criticality cannot be reached. Applying a burnup credit that uses realistic depleted concentrations of fissile radionuclides in the design of the repository can significantly reduce the cost of the repository and the transportation of spent fuel to the repository.

In any spent fuel disposal project there are universal repository programme constituents10 that need to be prepared:

• legal and regulatory framework

• organizational structure

• siting approach

• design concept

The following are basic technical principles that are applied for all well-developed disposal concepts:

• Encapsulation of spent fuel (or HLW) in a tight canister with a very long expected lifetime.

• Assurances that the conditions in the repository will allow the canister to remain intact and tight for as long as possible (such conditions include, for example, mechanical stability, stable geochemical conditions and very limited ground water movement, which could bring corrosive agents into contact with the canisters).

• Backfilling the repository with appropriate materials and locating it in geological media that, together with the backfill, strongly limit water movement and eventually waste movement when the integrity of the canisters finally breaks down.

An example of a spent fuel repository demonstrating all these principles is shown in Fig. 15.15, which is a Canadian project for spent fuel disposal managed by the Nuclear Waste Management Organization (NWMO) in Canada and which is also very similar to concepts in other countries.

Reprocessing operations make it possible to achieve a reduction in both the inventory of very long-lived radiotoxic radionuclides and the volume of high — level waste by extracting and recycling plutonium and uranium. Vitrification then conditions this waste into a disposable waste form. Partitioning and transmutation would further reduce the long-lived radiotoxicity of the waste by removing (partitioning) the minor actinides and then fissioning (transmuting) them by recycling in fast neutron reactors. The technical feasibility of transmutation of americium and neptunium has been established for oxide fuel in sodium-cooled fast reactors by the CEA (2004-9).

Other work by the CEA has developed aqueous solution processes such as PUREX to produce new separative processes. Trials have shown that at least 99% of the minor actinides can be recovered and a continuous process was established with a 15 kg inventory of spent fuel.

Once plutonium and uranium are recycled, partitioning of minor actinides and subsequent transmutation, would further reduce the high-level waste radio toxicity at 500 years by a factor of 10 to 100. This technology will require the deployment of a fast-neutron nuclear system and it is envisaged that development will continue as a complement to the new, fourth-generation fast reactors.

Countries that utilise all parts of the nuclear fuel cycle will generate a very much wider range of wastes. The UK, for example, is active in enrichment, fuel production, electricity generation and nuclear fuel reprocessing. Consequently the UK National Inventory of Radioactive Waste contains over 1000 different waste streams. These include everything from wastes that are barely above exemption level (a few becquerels per gram) to raffinate, a solution of fission products and minor actimides in nitric acid, which is the residue after uranium and plutonium have been extracted from spent nuclear fuel in reprocessing. Here the activity is so high that, when stored, it must be continuously cooled.* Reprocessing also produces long-lived intermediate-level wastes that require deep disposal. These include fuel element structural materials such as rod hulls and ends and process wastes such as ion exchange resins and ferric floc.

Wastes produced by enrichment and fuel fabrication mostly consist of items contaminated, respectively, with uranium or uranium and plutonium. A question hangs over the status of depleted uranium — the tailings from enrichment — and of reprocessed uranium, which is only slightly depleted but which may contain unwanted isotopes such as uranium-232 (see Chapter 16 for example). In the absence of a market for this material it may come to be considered as a waste.

Another significant bias in the LCOE concept is the assumption of a fixed electricity price. Nuclear plant may find it difficult to load-follow and its fuel makes a relatively small contribution to cost. Consequently, in seeking to maximise profit, the generator will maximise output, i. e. it will usually be in its interest to operate continuously and, therefore, at baseload. In this case the assumption of a fixed electricity price seems reasonable.

In contrast, let us imagine that, based on data like those shown in Fig. 5.1, a CCGT plant decides that it will aim to operate only when the electricity price is 50% above baseload price and that this results in the plant operating for four hours per day. This reduces the availability from 85% to 25% and increases the capital charges by a factor of 3.4 (85/25). Using the figures in Table 5.3 and assuming that other values remain constant, this increases the LCOE by about 33%. The difference between the 50% increase in the electricity price and the 33% increase in the LCOE represents increased profit.

By comparing LCOE values for nuclear and gas we are, in effect, viewing them as competitors. This calculation indicates that, on the contrary, they may in fact be complementary.

Enrichment is the process by which the proportion of 235U in natural uranium is increased above natural levels in preparation for nuclear fuel manufacture. The level of enrichment is typically expressed in terms of the proportion of 235U in the uranium, so that uranium compounds containing 4% 235U in relation to the total uranium is referred to as enriched to 4% or 4% enriched. This is also referred to as the assay or isotopic abundance. If enriched material is generated then some material with a level of 235U below natural must inevitably be generated as well. This material is known as depleted uranium (DU), with UF6 that has been depleted in 235U known as ‘tails’. The enrichment of uranium used for commercial power reactors is typically in the range of 3-5%.

The unit of measurement used to express the degree of enrichment applied through an enrichment process is the Separative Work Unit (SWU). It is related to the mass and enrichment level of the feed material, the product and the tails. The number of SWU in an enrichment process may be calculated using the equation:

W = P Vp + T Vt — F Vf [7.10]

where

W = Separative Work P = Mass of Product T = Mass of Tails F = Mass of Feed

Vp, Vt, Vf = Value function for the product, tails and feed respectively.

The value function is in turn given by the equation:

V = (1-2x).ln((1-x)/x) [7.11]

where

x = concentration of 2 35U in the material (product, tails or feed) expressed as a proportion, i. e. 0.0071, rather than 0.71%.

Despite being referred to as ‘work’, the SWU is not a true unit of energy (the dimension of the SWU is actually mass), although, for a given technology, there is a proportionality between the calculated SWU and the amount of energy used to achieve the desired separation. It requires relatively little energy to generate a small deviation from the feed concentration hence few SWU, while

large deviations require a great deal of energy and therefore a much larger number of SWU. Table 7.1 provides some examples of the relationship between SWU and product and tails enrichment.

The product enrichment is set by the requirements of the reactor, so that the amount of SWU required to generate that product is effectively determined by the extent to which the tails are depleted. The price of natural uranium therefore has a significant impact on the target tails depletion. The situation may be likened to squeezing juice from an orange, where the 235U may be viewed as the juice. It is relatively easy to get a small amount of juice from an orange, but to get nearly all of the juice requires a great deal more effort and it may well be easier to use more oranges! The concentration of 235U in tails material is typically in the range 0.2-0.3%. Table 7.2 provides some examples of the relationship between SWU, tails enrichment and the mass of feed material required.

The four main providers of commercial uranium enrichment services are:

1 Rosatom/JSC TVEL (Russia)

2 URENCO Enrichment Company Limited (UEC, UK/Germany/Netherlands/ USA)

3 AREVANC (France)

4 US Enrichment Corporation (USEC, USA)

Some capacity exists elsewhere in the world, notably in Japan and China, with the latter expected to become more prominent in the future particularly in its domestic market.

gas through it. The greater speed of the 235UF6 molecules means that they impact with the membrane more frequently and are therefore more likely to pass through one of the pores. The low pressure (product) side will therefore become slightly enriched in the lighter isotope while the high pressure (tails) side becomes slightly depleted. A continuous bleed on both product and tails side captures the enrichment. A diagram of a single diffuser unit is shown in Fig. 7.3.

The mass difference between the 235UF6 and 235UF6 is very small so that a single diffuser is capable of only a very modest enrichment, the theoretical maximum separation being 1.0043; in practice, the actual separation factor may be little more than half of this. This means that a great many diffusion stages are required to operate in series to give a product suitable for nuclear fuel manufacture. Furthermore, the tails material from many of the diffusers still contains commercially viable concentrations of 235U, so that many more stages are required to recycle the tails in a complex sequence of feed and re-feed so that only tails depleted to concentrations well below natural are discarded from the process. This interlinked series of stages is called a cascade and is illustrated in Fig. 7.4 . The diagram shows the principles of a cascade linked in both series and parallel, as is the case in centrifuge plants; diffusion plant cascades are simpler than this as they are linked in series only, with the throughput of each individual stage controlled by the size of the unit and the pressure applied. A diffusion plant contains thousands of stages, each of which contains a compressor, associated drive motor and a cooling unit in addition to the diffuser (the Oak Ridge Gaseous Diffusion Plant contained 5098 stages, although far fewer stages are required to enrich to commercial levels). The net result is a very large industrial facility; the Oak Ridge Gaseous Diffusion Plant was housed in the world’s largest industrial building when it was established in the 1940s, while the much smaller capacity Capenhurst plant was the largest industrial building in Europe under a single roof when built in the 1950s. Both plants occupied an area of close to 200 000 m2.

Although the principle of the technology is relatively simple, the design and manufacture of a membrane is no easy task. The membrane must be thin and must have a very small pore size, likely to be 20 nm or less in diameter. The number of pores must be maximised to keep the differential pressures as low as possible, yet inconsistencies in pore size must be minimised or performance will be compromised. The membrane must also be chemically resistant to UF6 and be of sufficient quality to perform reliably over many years. Materials that have been used or proposed for the membrane include nickel, aluminium oxide and fluorinated polymers.

A significant drawback of the gaseous diffusion process is its very high power consumption. The pressure drop across the membrane in each diffuser requires recompression of the gas before it is fed into the next unit. This continuous recompression requires a great deal of energy, with a commercial scale gaseous diffusion plant consuming as much electricity as a large city or small country; the French Eurodif plant consumes electricity at a rate of over 2000 MW when operating at full capacity, greater than the average 1800 MW electricity consumption rate for Wales in 2009. The recompression of the UF6 also generates a lot of excess heat, which must be removed from the system. This requires highly efficient cooling systems, which in turn require the use of refrigerant gases. Chlorofluorocarbon (CFC) gases such as Freon are highly efficient refrigerant gases and were a natural choice when gaseous diffusion plants were built. The
ozone depleting properties of CFCs are now known and their manufacture is banned under the Montreal Protocol, making operation of the old diffusion technology increasingly difficult.

It is a credit to the developers of gaseous diffusion technology that it has operated successfully for over 60 years and that during that time it has been responsible for producing the nuclear fuel required to power hundreds of power stations that, in turn, have provided electricity to many millions of homes and industrial premises. The technology has been under severe challenge from gas centrifuge enrichment for many years, however, and has now, finally, become economically obsolete.

The purpose of the fuel rod is to keep the fuel in a well-defined geometry and to provide the first barrier separating the fission products from the environment. Preserving its integrity is therefore the primary goal of fuel design and rules for reactor operation. As illustrated in Fig. 9.8, a fuel rod consists of a cladding tube

Lower end plug j — Fuel pellet j- Cladding

into which fuel pellets are inserted, a compression spring keeping the pellets axially in place during handling and transport, some extra space (plenum, about 10% of the fuel volume) to accommodate gaseous fission products and thermal expansion of the pellet stack, and end plugs welded onto the tube closing it and making it leakproof. The rod is filled with helium at a pressure depending on reactor type. Typical dimensions of cladding and fuel pellets are listed in Table 9.2.