During normal operation, because the BWR uses a direct cycle, the heat rejection route is via the turbine and condenser. Under fault conditions, to preserve the barriers to fission product release the main steam and main feed isolation valves are closed by the protection system. These are shown on Fig. 10.12. Core cooling must be continued and this, for non-LOCA conditions, is achieved by either the reactor core isolation cooling (RCIC) system or the isolation condenser system. The latter is used in older GE BWR3 plants, but because it is a passive system it is also used in the advanced passive plants (see 10.11).

The RCIC is shown schematically in Fig. 10.15. The RCIC consists of a steam turbine-driven pump and its associated pipework and valves. Using steam from the steam line to power the turbine pump, it draws water from the condensate storage tank and injects it into the main feed line. The system starts automatically on low water levels being detected in the vessel.

10.15 Schematic of a BWR reactor core isolation cooling system (Source: USNRC).

©

BWRs also have ECCSs to provide core cooling in the case of LOCAs. As is the case for PWRs this must cover a range of break sizes and so there are both high and low pressure systems. The high pressure coolant injection (HPCI) system is similar in configuration to the RCIC and is independent of ac power or external cooling systems. A turbine-driven pump injects make-up water from the condensate storage tank into the main feed line. The pump can supply water at pressures above the rated reactor pressure and will continue to be effective down to pressures at which the low pressure systems can operate. Excess steam will be discharged to the containment suppression pool, which acts as the ultimate heat sink in the short term.

In addition a high pressure core spray (HPCS) system is provided. This uses electric pumps supplied by the diesel-backed essential electrical system, which draws water from either the condensate storage tank or the suppression pool and sprays it onto the top of the core using spray rings mounted in the upper part of the core barrel above the top of the core.

To provide an alternative, should the high pressure systems be unavailable or unable to recover the water level, an automatic depressurisation system (ADS) is provided. This opens selected safety relief valves to depressurise the RCS by discharging to the suppression pool. This then allows the use of the low pressure ECCSs, which are illustrated in Fig. 10.16.

The low pressure ECCS system consists of two independent subsystems: the core spray system and the low pressure coolant injection (LPCI) system. As is the case for PWRs the LPCI system is also used for residual heat removal. The core spray system consists of a number of redundant loops consisting of a diesel — backed electrically driven pump, which draws water from the suppression pool and sprays it via the low pressure core spray sparge ring, which is mounted just below the HPCS sparge ring.

The RHR pumps in their LPCI mode perform three main safety-related functions: LPCI (restore and maintain RPV water level when the RPV is depressurised); containment spray (condenses steam and reduces airborne activity) and suppression pool cooling (provides an external ultimate heat sink).

The six reactor types proposed by the Generation IV forum include developments of the HTR thermal reactor and the high-temperature gas-cooled fast reactor (using pebble or prismatic fuel elements as well as plate designs). The object of the GenlV designs includes development of systems giving high utilisation of fuel, reduced waste arisings and the potential for burning actinides. Development is scheduled to progress through the second and third decades of the twenty-first century.

There have been many variations proposed on the basic gas-cooled reactor. Designs have been studied for helium, or helium-xenon coolant reactors for use in space exploration. Adams Atomic Engines of the USA (now defunct) proposed designs for very small-to-medium reactors cooled with nitrogen. In the UK, designs were studied for CO2 coolant reactors loaded with graphite dust to enhance heat transfer performance, and a design for a high-temperature variant on the AGR using helium coolant (the HTAGR) was extant during the late 1970s (anticipating one of the current GenIV designs).

There are three possible mechanisms for clad wall thinning due to the action of the coolant: oxidation (i. e. the chemical reaction of the cladding material with an oxygen-bearing coolant, or with oxygen dissolved in the coolant), erosion (i. e. the wearing away of the metal due to the forces induced by the flowing coolant) and dissolution (i. e. leaching of the cladding material). Oxidation and dissolution are also commonly known as corrosion. Excessive wall thinning can lead to failure of the cladding.

Which of the three wall thinning mechanisms dominates depends upon the coolant type, the cladding material, the coolant dissolved oxygen concentration, the clad surface temperature and the coolant velocity. Only oxidation is pertinent to light or heavy water coolants; dissolution is of primary importance in sodium; dissolution and erosion are both significant in lead or a lead-bismuth eutectic. Oxidation, erosion and dissolution can all potentially be reduced or eliminated by applying a suitable coating to the cladding surface, or by surface treatment. Erosion can also be prevented by limiting the coolant velocity. If selective dissolution of cladding constituents occurs the cladding may become embrittled.

The extent of corrosion is very much dependent upon the cladding material and the coolant type. The two most common combinations are austenitic steel cladding in sodium, and zirconium alloy cladding in water. The corrosion is generally limited in the case of the former, with 20 to 30 ^m of the cladding wall dissolved away (Bailly et al., 1999). In the case of the latter, the extent is more variable and is dependent upon the exact cladding alloy. With respect to Zircaloy-2 and Zircaloy-4, the corrosion can be considerable, with an oxide thickness of ~100 ^m (above this thickness oxide spalling tends to occur), although on average the oxide thickness tends to be significantly less than this — approximately 50 ^m for a PWR Zircaloy-4 clad pin with a typical pin average burnup of 45 MWd/kgU (Chapin et al., 2009). The oxide thicknesses are generally significantly less with more modern cladding alloys. Corrosion tends to be uniform in PWRs and CANDU reactors, but may be nodular in BWRs.

The clad corrosion considered above pertains to corrosion of the outside of the cladding by the coolant. However, corrosion of the inside of the cladding may also occur due to chemical reactions between the cladding, the free oxygen in the fuel (liberated by fission in oxide fuel), and the fission products. The potential for corrosion by free oxygen in oxide fuel is enhanced by oxygen migration (see 14.2.10). In the case of LWR fuels, where discharge burnups are moderate, clad temperatures are relatively low and a zirconium alloy cladding is used, oxidation of the clad inner wall is limited. However, for fast reactor fuels, where burnups and clad temperatures are considerably higher and a steel cladding is used, corrosion of the clad inner wall is more extensive.

There are three distinct types of clad internal corrosion that occur in fast reactors: early-in-life, ROG and RIFF (Bailly et al., 1999). Early-in-life corrosion is localised and occurs only at high powers in the first ten days of irradiation, but can cause clad penetration. It is caused by free iodine and tellurium (which after ten days have mostly reacted with rubidium and caesium to form non-corrosive compounds). This type of corrosion can be prevented by limiting the pin powers during the first ten days of irradiation. ROG corrosion (named after the French term — reaction oxyde gaine) is a general fuel-cladding reaction, and is the main corrosion type of interest. RIFF corrosion (named after the French term — reaction a I’interface fissile-fertile) can occur at the interface between the fissile and fertile fuel in a pin with axial blankets. Both are due to complex chemical reactions between the free oxygen in the fuel, the fission products (principally tellurium) and the metallic elements in the cladding (principally iron, chromium and nickel). The ROG corrosion rate is mainly dependent upon burnup and cladding strain. RIFF corrosion is less well understood.

Gaseous effluents come from process equipment and, to a lesser extent, from the workshops’ ventilation systems. Gaseous fission products are mostly collected from the shearing process and after dissolution of the used fuel assemblies and then undergo successive treatments. For example, almost all the tritium is trapped as tritiated water with the remainder released as tritiated steam. Iodine-129 is absorbed on element-specific filters made of zeolite and silver nitrate. Particulates are stopped by high-efficiency filters (99.9% efficiency). Minor residual releases are mainly ruthenium and krypton-85, which have an extremely small radiological impact.

The main part of the gaseous effluents is released through stacks that are high enough (about 100 m) to provide sufficient atmospheric dispersion to reduce the impact to acceptable levels. Flow rate and activity are continuously monitored through automatic systems and some other controls are performed through sampling and laboratory measurements. Independent checks on these measurements are performed by the safety authorities.

As an example, gaseous releases at the La Hague in France measured in 2002 are shown in Table 16.2.

For the first point, a simple indicator of the capture to fission balance is the reaction ratio a = Capture/Fission. This ratio is heavily dependent on the neutron spectrum, as can be seen in Table 17.2. For most TRUs, the most favorable (i. e. low) a ratio values are obtained in a fast neutron spectrum. The fact that MA isotopes act as ‘neutronic poisons’ (i. e. they have high a values) in a thermal neutron spectrum, indicates in principle that additional fissile enrichment is likely to be needed, if these isotopes are loaded into a thermal neutron core. In addition the table illustrates the well-known saying that the even-numbered isotopes of plutonium are not much fissionable and shows that this only applies to thermal reactors.

For the second point, a fast neutron spectrum reactor leads to fewer high mass isotopes since, compared to a thermal reactor, the TRU isotopes are more likely to fission. A typical example is the build-up of Cf-252 (a very strong neutron emitter by spontaneous fission), see Fig. 17.2 . A consequence of higher quantities of higher mass TRU isotopes in thermal reactor fuel (especially the shorter-lived Am and Cm isotopes, see Table 17.1) is an increase of the decay heat.

17.2 Cf-252 inventory in the core. (a) For grouped TRU multi­recycling in a LWR. (b) For grouped TRU multi-recycling in a FR.

A radionuclide-specific issue, whatever the neutron spectrum, is related to fuels containing large amounts of Am-241. Here, Cm-242 produced by neutron capture decays by a emission to Pu-238. a production leads to helium deposition in the fuel and this is sufficiently great to require special design measures to accommodate it.

For the third point, for a full understanding of the transmutation potential of different neutron fields one can use the very simple notion of total neutron consumption per fission, Dj, of an isotope family j, similar to that introduced in Ref. 4. In the case of a negative D (i. e., when the j-family produces more neutrons than it consumes), a core fuelled by j-nuclides can produce enough neutrons to destroy the source material at equilibrium if the neutron excess compensates for parasitic captures (e. g. by structures, fission products, etc.) and for neutron leakage. In the case of a positive Df neutron consumption by the j-nuclides dominates over neutron production and the core cannot support transmutation unless a supplementary neutron source is provided.

The global neutron balance of a core must consider the total neutron production (or consumption) of the relevant fuel families, the parasitic captures of other core components (C ) and of the accumulated fission products (CFP) and neutron leakage (Lcore). A very simple and general equation for the neutron surplus (NScore) can be written as:

— D) — C — CFP — L = NS [17.1]

fuel par FP core core

with all quantities of Eq. 17.1 normalized to one fission. For a critical system, the neutron surplus must equal zero and the neutron production of the fuel feed must be sufficient to overcome the other loss terms. Thus, D) , is a useful parameter for quantifying the transmutation potential of a given isotopic fuel mix.

Due to the fact that most MAs act as strong neutron ‘consumers’ in a thermal neutron spectrum (i. e. having unfavorable fission-to-capture rates, see Table 17.2), the D values corresponding to a LWR fuel loaded with MAs can become positive, indicating that the neutron balance is very tight (i. e. no ‘neutron excess’ is available) and that a further fissile enrichment would be needed (a potentially significant economic penalty). An example of NScore values is given in Table 17.3.

Where a repository is already in operation, waste generators will usually be anxious to process and despatch their low — and intermediate-level waste as soon as possible — recognising that the business of an NPP is electricity generation not waste management. Where there is no disposal facility in existence, however, (and this is usually the case for deep disposal) there may be considerable uncertainty about how the waste ought to be processed to meet the waste acceptance criteria of the yet-to-be-created facility. In such circumstances the waste producer may prefer a ‘wait and see’ strategy in which the waste is kept in its raw state. This may not be acceptable to the regulator, however, who, understandably, will want to see that the waste is stored in a passive state. Preliminary treatment is one way around this dilemma but has the significant disadvantage that the waste will then need to be processed twice: once for storage and again for disposal. A solution that has been used successfully in many countries is the formulation of ‘preliminary waste acceptance criteria’, basing these upon an outline design for a waste disposal facility that is considered to be both feasible and affordable. Waste acceptance criteria (WAC) are normally established by the body that has ultimate responsibility for radioactive waste disposal: the waste management organisation (see next section).

Financing, the second component of the capital cost, is derived from the overnight costs (discussed above), construction time and the interest rate (approximated here by the discount rate). Programme delays from whatever source — engineering, regulator intervention or public opposition — will damage the project by increasing the financing costs. In all circumstances, expanded experience of NPP construction is likely to be beneficial. With respect to the engineering work, control and mitigation of commercial risk is achieved via the contractual arrangements and effective project management. Contracts should aim to allocate risk to the party that is best able to control it. This needs to be done at the outset because it will be difficult to change later. The duty of controlling construction delays, for instance, will normally be allocated to the main contractor who may, of course, redistribute this amongst his sub-contractors. It is usual to see fixed price contracts for construction with incentives and penalties linked to the project schedule.

A recent development is the pre-licensing of NPP designs. This aims to address any obstacles to licensing before making any investments in plant or on-site work and may be seen as a way of mitigating regulatory risk. Another method is phased financing, which recognises that the risks attendant on the various construction stages are of varying magnitude and may merit different treatment in terms of project management, method of financing, allocation of risk and mitigation measures. Financing is therefore arranged separately for each phase so as to recognise the differing circumstances. Risk mitigation may also be effected by seeking equity partners to spread the risk. These could include owners, vendors, government, banks and, as in the case of the new plant in Finland, customers.15

Unexpected increases in interest rates and inflation during the construction period may also cause concern: the first because it directly affects financing costs and the second because of its potential to reduce real returns. Finally, many NPP components are likely to be imported from abroad and a significant change in currency exchange rate of the importing country will also cause capital costs to increase.

The established commercial enrichment processes of gaseous diffusion and gas centrifuge enrichment both generate depleted tails material in the form of UF6 , with the quantity of tails material significantly greater than the quantity of product. The tails material still contains potentially valuable 235U, which is worth recovering should the price of natural uranium reach sufficiently high levels and industry preference is therefore to store tails material as UF6. The volatility and chemical reactivity of UF6 means that it is not ideal for long-term storage, however, and the quantities involved require a lot of storage capacity, so that conversion to a more compact and stable form must at some point be considered. The preferred chemical form for long-term uranium storage is U3O8, the same material that is fed into the conversion process earlier in the nuclear fuel cycle. The process of converting UF6 back into U3O8 for long-term storage is therefore known as deconversion.

A large deconversion plant has been operating successfully at Pierrelatte in France since the 1980s, with over 300 000 tonnes of UF„ of tails deconverted. The facility is owned and operated by AREVA. The process strips fluorine from the UF6 in the form of HF using steam and hydrogen to give a U3O8 product. The HF is collected as concentrated hydrofluoric acid and sold into the commercial market as a high-grade feed material for chemical processing. The chemical process is shown in Fig. 7.7.

In the first stage of the process, UF6 is mixed with superheated steam in a hydrolysis chamber, generating solid UO2F2 and HF gas via a gas phase reaction (see Eq. 7.1). The reaction takes place at 250-300 °C. The steam is added in excess, which allows the HF and excess steam to be condensed and collected as concentrated hydrofluoric acid, with the amount of excess steam driving the concentration of the acid.

In the second stage of the process, the UO2F2 is reacted with more superheated steam at a higher temperature and with some hydrogen present giving the overall reaction:

3UO2F2 + 2H2O + H2 ^ U3O8 + 6HF [7.12]

The plant at Pierrelatte uses a rotary kiln for the pyrohydrolysis with a reaction temperature of around 750 °C. The steam is again added in excess to allow condensation and collection of the HF as hydrofluoric acid. The process is designed so that the hydrofluoric acid output from the process, including the effluent from any wet scrubbing systems, is generated at a suitable concentration for onward commercial use with appropriate, standard concentrations being 40%, 60% or 70% by weight. The U3O8 is generated as a powder, which is collected and stored.

A number of new deconversion facilities have recently been, or are in the process of being, constructed. A plant has been built in Russia based upon the Pierrelatte design and another is in the course of construction in the UK. Further plants have been built at Paducah and Portsmouth in the USA to process tails material from the US military programme. These are of a slightly different design as a single fluidised bed reactor is used, rather than a hydrolysis chamber and kiln, but the chemical process of reacting the UF6 with a mixture of steam and hydrogen is essentially the same.

A few other types of fuel have been tried on a minor scale or have been irradiated in research reactors or as lead rods in commercial nuclear power stations. Metal fuel is still used in the UK Magnox reactors, but this reactor type is being phased out and will soon disappear. Metal fuel is produced from UF4 , which is mixed with magnesium metal and heated to about 600 °C. The UF4 and magnesium react leaving molten uranium metal at the bottom of the furnace. After cooling, the uranium metal is melted again, cast into rods, which are machined to size and length, and put into magnesium alloy tubes.

Vibration-compacted (VIPAC) fuel is made of granules with different size fractions, which are directly filled into the fuel tube without first making pellets. Vibration compaction can achieve a smear density of about 80-85% of theoretical density. Experimental irradiations have shown an acceptable performance (Knudsen et al, 1977; van der Linde, 1982).

After burning and unloading, spent LWR fuel has a higher equivalent fissile content of U-235 than natural uranium; it also contains plutonium. Fuel for ‘Direct Use of Pressurized Water Reactor Spent Fuel in CANDU’, DUPIC, could therefore continue irradiation in CANDU reactors after refabrication. In the DUPIC fabrication process, uranium, plutonium, fission products and minor actinides are retained in the fuel powder from which new pellets are produced (Bae et al., 1998; Lee et al., 2008). DUPIC fuel has to be fabricated in shielded facilities.

The natural uranium fuel cycle has worked well in the CANDU reactor: both front — and back-end of the fuel cycle are economical, fuel performance has been excellent and uranium utilization is high. The advantages of introducing an advanced fuel cycle into the CANDU reactor would need to be compelling in order to offset the costs of fuel qualification, increased fuel costs and the analysis costs for reactor safety and licensing. The considerations would include not only economics, but local, national and strategic aspects such as security, diversity and availability of energy and fuel supply. The features that enable the use of natural uranium fuel facilitate the use of other advanced fuel cycles such as RU, LEU, MOX, minor actinides and a variety of thorium fuel cycles.

Before describing specific fuel cycles, some general characteristics will be listed: [22]

• The simple, small fuel bundle design facilitates remote processing and fabrication for highly radioactive recycle fuels.

• On-line refuelling in a pressure tube reactor provides flexibility in fuel management to accommodate both high and low reactivity fuel and to shape the axial and radial power profiles.

• An extensive array of flux detectors in the core ensures knowledge of the flux and power distributions in the core, regardless of fuel type.

LEU

The optimal fuel enrichment in the CANDU reactor from the perspective of uranium utilization is around 1.2% U-235, which would result in a burnup of ~21 MWd/kg HE and uranium utilization almost double that of an LWR (Boczar et al., 1996). This enrichment could be accommodated by a 2-bundle shift, bi-directional refuelling scheme (Younis and Boczar, 1989a). The economics of the use of LEU in CANDU will depend of course on whether the increased costs for enrichment, conversion, fuel fabrication, fuel qualification and safety and licensing are offset by the higher fuel burnup. Enrichment can also be used in the CANDU reactor to tailor reactivity coefficients. For instance, in the Low Void Reactivity Fuel (LVRF) bundle, a BNA material in the centre element of the fuel is compensated by enriched uranium in the outer elements (Boczar et al., 1992). BNA content and enrichment can be chosen to give desired values of coolant void reactivity and fuel burnup. This of course is at the expense of fuel utilization. Enrichment can also be used to achieve other objectives, such as power uprating by flattening the radial power distributions across the core (without exceeding maximum bundle or channel powers).