6.6.1 Boilers

The primary function of the boilers is to transform the heat from the reactor into a form which is suit­able for electrical power generation. In this respect the AGR represents a major advance from the magnox reactors in that the boilers are designed specifically to provide steam to 660 MW turbines of exactly the same design as those used on the modern fossil-fired stations. The boiler temperatures and pressures are therefore higher than on the magnox stations and in addition there is an integral gas-heated reheater. The reactor itself has a heat output of 1550 MW and delivers carbon dioxide gas at 6l5°C to the boilers, which raise 500 kg of steam per second at 54I°C and 166 bar for expansion in the turbine HP cylinder and subsequently reheat it to 539°C at 40.7 bar for expansion in the IP cylinder. This duty requires 300 kilometres of boiler tubing in each reactor and the whole boiler system accounts for some 10% of the capital cost of the station.

Apart from its function in transferring power, the main boiler system plays a major role in the safety of the reactor in that it forms part of the primary cir­cuit pressure boundary and is the principal means of removing the heat produced within the reactor vessel post trip. An additional and much smaller decay heat boiler is installed at the base of the main boiler. This provides the essential diverse post-trip cooling system and is served by completely independent feed and steam systems.

The boiler systems in the various AGRs are all similar in principle, but two basically different types of boiler tube bank configuration have been devel­oped and are in service: the serpentine boiler shown in Fig 2.88 and the helical pod boiler shown in Fig 2.89. The description which follows applies in general to both types of boiler but in detail primarily to the serpentine type of boiler as installed at Heysham 2, Hinkley Point В and, broadly, Dungeness B. The helical pod type of boiler is installed in the Hartlepool and Heysham l reactors.

The boilers are located within the reactor pressure vessel in an annulus formed between the gas baffle and the reactor pressure vessel wall. This is a de­velopment of the design principles originally estab­lished for Oldbury and Wylfa magnox stations. The boilers are of the once-through subcritical type, chosen in order to minimise the number of reactor pressure vessel penetrations, and a single stage reheater bank is located above each high pressure unit. The boiler annulus is partitioned into four quadrants, each con­taining three rectangular boiler units and two gas circulators. Each quadrant is operated separately and any one may be taken out of service while the reactor is on load. Each of the twelve individual boiler units (eight for helical pod boilers), weighing 120 tonnes, is in a 16 m tall vertical rectangular case, containing banks of closely packed boiler tubes. The tubing is arranged in horizontal rows with vertical serpentine return bends; the horizontal tube axes are arranged in the vessel circumferential direction to maximise the length of straight tubing.

Hot gas from the reactor core is drawn down over the banks of boiler tubing by the gas circulators lo­cated underneath the boilers at the bottom of the annulus. The gas is constrained to flow down over the banks of boiler tubing by means of the casings which surround each boiler unit and by the gas seals which form a floor across the annulus around the base of the boiler casings just above the circulators. The feedwater connections to the boilers pass through penetrations in the walls of the concrete pressure vessel and enter the unit casing from below, while the superheated steam outlet connections pass through penetrations near the top of the casing.

The serpentine tubes, which make up the decay heat

boiler and the main economiser/evaporator/superheater

sections, are assembled in pairs by means of welded spacers to form vertical platens. Forty-four such pla­tens hang side-by-side in the module casing and are supported at regular intervals across and down the module by links attached to beams which span the casing. That part of the module which comprises these platens and their stainless steel casing is supported at its base on mild steel beams. The beams are slung from flexible links anchored to the gas baffle skirt on the one side and to the reactor pressure vessel wall on the other. The load is transferred from the module casing to the beam through a pin joint. This permits the anchor points of the beam to move rela­tive to one another without themselves causing the module to tilt, whilst at the same time accommodat­ing the slight radial tilt which is produced by thermal expansion of the superheater tailpipes.

The reheater section of the module is located above the superheater section and is hung from the top slab of the reactor pressure vessel. The tubing is again arranged in the form of platens, but in this case there are 36 platens each of which consists of four parallel steam paths in the same vertical plane. The casing can move relative to that of the lower part of the module and a gas-tight seal is provided by means of a flexible foil.

The tube materials are chosen to avoid excessive gas or waterside corrosion. They are therefore graded from lCr^Mo in the economiser, through 9Сг 1 Mo, to 316H austenitic steel for the superheater and reheater where the gas temperatures are high. The роіпГ of transition from the 9CrIMo ferritic steel to the austenitic material is particularly sensitive. It Js necessary to ensure that the gas temperature does not exceed 550?C in order to avoid excessive gas-side oxidation of the ferritic steel, and that the steam is superheated to a temperature sufficiently above sat­uration temperature to avoid any risk of stress cor­rosion in the austenitic steel. The junction between the two materials is achieved with a short transition section of Inconel 600, which has a coefficient of expansion roughly halfway between those of the fer­ritic and austenitic tubing materials.

The boilers are designed to withstand earthquakes up to a specified severity. Vertical seismic forces are accommodated by the main support beams; radial forces by the superheater penetration sheath tube connections to the upper casing and to neighbouring units via the annular ring at the base of the casing; and circumferential forces by special upper and lower seismic restraints which link the unit casing to dummy penetrations in the vessel wall.

The flow/pressure drop characteristics of once- through boilers can cause large flow and tempera­ture differences to occur between tubes and between boiler units if the design is not undertaken carefully. The temperature differentials would be undesirable

REHEATER-MAIN BOILER GAS SEAL

SUPERHEATER TAIL PIPES

SUPERHEATER SUBHEADERS

AUSTENITIC STAINLESS-STEEL SECONDARY SUPERHEATER tubes

SPECIMEN ACCESS 1 TV ACCESS TUBE

Fic. 2.88 Serpentine main boiler and reheater unit

because they would lead to thermal stress or even to the carryover of water into the turbine. The explana­tion of this potential flow instability is as follows. If the water flow in a boiler tube is increased the pressure drop will increase in proportion to the square of the mass flowrate (Fig 2.90 (a)). However, the length of tube occupied by water will also increase relative to that occupied by steam. This will increase
the gravitational pressure drop but, since the pressure drop per unit length is greater in the superheater, there will also be a reduction in pressure drop. The fiow/pressure drop characteristic may therefore be cubic (Fig 2.90 (b)) which means that for certain pressure drops three different flowrates may exist. In addition to this ‘static’ instability, it is possible for a periodic oscillation of the flow to occur in any one

tube and this is usually coupled out of phase with other tubes.

In the AGR boilers, large numbers of tubes are connected in parallel between common feed and steam headers and would therefore operate at substantially different water flowrates and temperatures despite the fact that they all have the same pressure drop. These problems are controlled by the installation of an ori­fice at the inlet to each boiler tube such that the total pressure drop/flowrate curve has a continuously positive gradient (line C in Fig 2.90 (b)). However, the small size of orifice which is required to provide control at low loads would lead to prohibitively high pumping costs at full load. A compromise is there­fore accepted in which control is provided at full power but some instabilities are tolerated for brief periods during start-up and post trip. A complex suite of computer hydrodynamic models is used to predict the steady state, static/dynamic stability and transient responses of the boilers.

Feedwater passes through the boilers at the rate of half a tonne per second throughout most of their 30 year working life. However, the tube materials are made of ordinary steels which are actually corroded in this environment. The steel is in fact protected only because the initial corrosion products form a thin surface film which effectively inhibits further attack. The film is magnetite (FejC>4) and is virtually insoluble in pure water. But if contaminant solutes or oxygen are allowed to enter the boiler, very high local concentrations of acids, alkalis or salts will be generated at the boiling dryout zones. Under these conditions the magnetite film becomes much more soluble and breaks down. Corrosion then proceeds rapidly and unchecked because the magnetite produced by this on-load corrosion is porous and unprotective. The austenitic steels in the superheater are susceptible to ‘stress’ corrosion in these same aggressive solution environments, particularly in the presence of oxygen. Whilst short-term wetting of the austenitic is accept­able if the water quality is good, a superheat margin is normally maintained at the 9Cr-316 transition joints to ensure that the austenitic sections run under dry steam conditions.

In regions of high water velocity, a tube metal wast­age phenomenon known as ‘erosion-corrosion’ may occur and can lead to extremely high rates of metal loss. This process is a function of velocity, pH, temperature, oxygen concentration and the chromium content of the tube metal. Severe damage due to this corrosion mechanism has been experienced on both magnox and AGR stations in the past. The boiler tube inlet orifices and the bends in the economiser are the most susceptible areas. However, the phenomenon is now better understood and is controlled by raising the pH and/or oxygen dosing. The inlet orifices and erosion sites downstream are made of stainless steel and, in the latest stations, 1 Cry Mo is used in place of mild steel for the economiser and decay heat boiler tubing in early designs.

The carbon dioxide gas surrounding the boiler can be equally aggressive and, if gas temperatures are too high, may cause gross metal loss of the ferritic boiler steels through an accelerated breakaway type of oxidation. Mild steel had been used for the whole boiler in the magnox stations, but this has a tem­perature limitation of 350°C due to oxidation in the

Fig. 2.90 Flow/pressure drop characteristics of once-through boilers

carbon dioxide. This material (or ІСгтМо) is there­fore used only for the economiser and the decay heat boiler in the AGRs. Austenitic stainless steel is suit­able for all surfaces exposed above this temperature and is therefore used for the reheater and superheater which are exposed to the hottest gas at up to 615°C. However, because of the risk of stress corrosion in the stainless steel if any wetness is present in the steam, high chrome ferritic steel 9CrlMo is chosen for the larger, intermediate region of the boiler where evapora­tion and the primary stage of superheating takes place. Gas side oxidation of this material is acceptable up to about 550°C, beyond which the oxide weight gain induces breakaway and the oxide layer is no longer protective. The corrosion rate then increases by one or two orders of magnitude and would lead to pre­mature failure of tube supports. Since there are major power output benefits in running at higher gas tem­perature, the boilers are run as close as possible to this constraint. The judgement is based on very exten­sive oxidation research programmes and upon detailed computation of boiler temperatures and gas mixing.

The gas flow through the boilers has a high energy and density and is extremely noisy. The boiler tubes,

which are elastically suspended across this fluid flow, are liable to be excited into vibration via a number of ~ different fluid elastic mechanisms. This may lead to fretting damage at clamped tube supports or to fatigue damage at welded tube supports. This is a common problem in shell and tube heat exchangers, but is a cause for particularly careful scrutiny in the design of the AGR boilers because the extremely limited access allows virtually no repair or modification during the 30 year station life. The main excitation phenomena are vortex shedding, fluid elastic whirling, turbulent buffeting and acoustic resonance. However, these are complex phenomena and are not yet fully understood despite many years of theoretical and experimental investigation.

In vortex shedding, inwardly spiralling vortices are shed into the wake alternately from one side of the tube and then the other. An alternating counter­circulation is induced around the tube and this pro­duces a fluctuating side thrust (aerodynamic lift) and also a fluctuating drag force. The transverse force is several times greater than the stream-wise force. The frequency with which vortices are shed is characterised by the Strouhal number:

The Strouhal number has a value of 0.2 for an iso­lated tube but is different for a closely packed tube bank (as is the true nature of the wake) and is there­fore normally determined from tests on rigid tube banks. If the vortex shedding frequency coincides with the natural response frequency of the boiler tubes, then coupling occurs and the tube motion feedback enhances the driving force such that unacceptably large vibration amplitudes are induced. Once coupling has occurred, the vortex shedding frequency ‘locks-in’ to the tube frequency and coupling persists over a broad velocity range (about ±30%). Experience has shown that, in order to avoid coupling, the designer should ensure that the natural response frequency of the boiler tubes is about three times higher than the vortex shedding frequency.

Fluid elastic whirling is a vibrational instability phe­nomenon which can arise in a steady fluid flew and which is in fact more common than vortex shedding resonance. When a tube within an array is displaced, the aerodynamic pressure forces around the tube are altered in such a way as to induce movement of the tube. If the tube damping is relatively low, such that it is unable to dissipate the fluid energy being sup­plied, the tube motion will increase and so will the fluid forces. This divergent instability will occur at the natural response frequency of the tube and is characterised by an elliptical whirling tube motion which is coupled out-of-phase with the similar motion of adjacent tubes in the array. A guide to a satisfac­tory relationship between fluid elastic excitation and tube damping has been obtained by Connors:

V/fD < K(m6/eD2)T

where V = is the maximum velocity between tubes

F = is the tube lowest natural frequency

D = is the tube diameter

M = is the tube mass plus added mass of entrained fluid per unit length

Turbulent buffeting may be regarded as a decayed and incoherent form of vortex excitation, generally in­volving much smaller amplitudes over a much broader frequency spectrum. For these reasons it rarely presents a problem, provided that the structural natural response frequency and the prospective vortex shedding frequency are widely separated.

Apart from these examples of potential fluid force excitation of body resonance, if the boiler casing dimen­sions are such that the frequency of the natural acoustic standing wave coincides with the vortex shedding fre­quency, then it is possible for the sound to become amplified to the extent that it can cause damage to the steam generator structures. It is for this reason that the AGR boiler units are each divided in half by a vertical acoustic baffle division plate.

In the case of the Heysham 2 AGR, the reactor gas instrument pipework varies in size between 3 mm and 22 mm outside diameter with larger sections of pipe being used for blowdown manifolds. The pipework is solid-drawn cold-finished seamless stainless steel to BS3605 and is to grades 316, 321 and 359 S18. Over the outside diameter pipe range between 6 mm and 22 mm, joints are made by use of socket welds with smaller pipes having silver alloy brazed sockets. Any pipes above the: range are butt welded. Some instru­ment connections are made by means of compression fittings.

All joints between the reactor penetration and the primary isolation valve of every circuit, and ten per cent of all others are subjected to surface crack de­tection. The examination is made by the use of a low halogen red dye penetrant process HF-P, HF-R and HF-D performed by qualified personnel, thus meeting the requirements of the ASME boiler and pressure vessel code Section 5, Article 6.

The pipework and associated components are built to a design pressure of 60 bar(g) and are pressure tested at 52.5 bar(g).

All pipework external to the reactor penetration up to and including the primary isolation valves is seismically qualified by analysis, as are the structures to which the pipework is attached.

The primary isolation valve for the burst can de­tection sample pipes forms part of the fuel standpipe closure unit and the 3 mm pipes, which are external to this, are clipped to the side of the pile cap upper cable trunking for routing off the pile cap. In addi-

lion to the self-sealing coupling, there is a secondary isolation valve at the instrument rack.

To prevent precipitation of tritiated water in the instrument pipework, it is maintained at a temperature well above the precipitation point associated with the reactor pressure by means of trace heating and lag­ging. At certain controlled points the temperature is allowed to reduce to enable any moisture that is m the gas to precipitate out and be collected and drained. This system of extraction is provided to remove excess moisture during, say, a boiler leak and not to remove the normal level of moisture that exists m the reactor gas.

halves are, in general, the preferred bellows seal vale with a secondary gland seal. They are stainless ясе! with socket weld connections.

The reactor coolant temperature is maintained at its demanded value during normal operational transients

by regulation of the insertion of the appropriate

banks of control rods.

The consent to operate a reactor is given for two year periods. Within a maximum of two years since the licensee was given consent to start it up, it must be shut down for inspection. The object of the in­spection is to give sufficient confidence and assurance that the reactor can run a further two years in a safe and untroubled way.

The inspection looks at all fixtures and fittings that are both inside and outside the reactor pressure vessel. The following is a list showing the type of compo­nents inspected, although the list may vary slightly from location to location: [31]

ness and corrosion, and any signs of failure, es­pecially welds.

• Use of NDT techniques to look at the soundness of welds, especially those associated with pressure vessel penetrations.

• Inspection of any moving part or link within the pressure vessel for freedom of movement and cor­rosion.

• Take samples of graphite core for scientific analysis to look at its condition and deterioration,

• Check fuel channels and control rod channels for straightness to identify any distortion of the core,

• Check the soundness of pressure vessel thermal insulation in the case of concrete pressure vessels.

• Check the insertion rates of entry of control rods under free-fall in a hot reactor core.

• Check the soundness of the pressure vessel cooling system to ensure there are no internal leaks to the concrete pressure vessels.

• Inspect the soundness of gas circulator or boiler gas side isolating valves and ensure they function correctly.

• Check boiler structures and inspect for any boiler leaks that may have been occurring.

• Carry out a debris check in the bottom of boilers and be satisfied that debris is as expected.

This list represents the main reasons for the biennial overhaul, but others may be included at the request of the Nuclear Installations Inspectorate.

Following the inspection, a meeting with the Nil will be held to satisfy them it everything is in order and a further two year period of operation can occur without any major failure or problem. At this meet­ing the results of the findings of both the biennial inspection and the maintenance schedule results will be tabled. There will also be discussions on future work and the expectations of the conclusions from any test carried out. Providing the Inspectorate is satisfied, the chief inspector will issue a certificate to the licensee to give permission for the reactor to be started up. During the process of the biennial overhaul, the inspectors will have been watching the results of all activities and satisfying themselves as the work proceeds.

Within 14 days of the start-up certificate being given, the licensee must have presented a report on the result of the inspection together with all technical data of inspections and work carried out. If any test has indicated that a problem exists remedial work must be identified or it must be shown that the defect has no material effect on safety of the component.

The foregoing delayed neutron data and calculations have been based on thermal fission of U-235. Other fissile materials have different delayed neutron char­acteristics, for example, Table 3.3 gives the yield and mean life values (for the 6-group model) in the case of thermal fission of Pu-239.

Note from Table 3.3 that the total delayed neu­tron yield from Pu-239 is significantly less than that from U-235. Thus as the isotopic content of the fuel in a power reactor changes with the irradiation, the time-dependent behaviour of the neutron population also changes. For example, the doubling times in an AGR on an equilibrium fuel cycle at current discharge irradiations are approximately 20% shorter than at start of life,

4.3.2 Shutdown kinetics

The shutdown state is a special case in reactor kine­tics. In order to provide a reading on the neutron flux instrumentation during shutdown and start-up, so that the operator is not ‘flying blind’, a neutron source is located in the reactor core. A typical neu­tron source is antimony/beryllium, Sb-124 being a ~r emitter and Be-10 responding with a (7,n) reaction (see Chapter 1, Section 3).

Consider a shutdown reactor with a given value of multiplication constant k, a typical value being 0.97. Suppose the neutron population is zero when a neutron source is suddenly introduced into the core. Consider in small timesteps the time-dependent be­haviour of the neutron population. Let the neutron source emit Q neutrons in each timestep. In each successive timestep, the neutron population will be multiplied by the multiplication constant к and an­other Q neutrons will be added by the source, so the neutron population will change from timestep to timestep as follows;

Q

Qk + Q — Q (k + 1)

(Qk + Q, к + Q = Q (к2 + к + 1) etc.

The value of к is less than unity, therefore the neu­tron population will build up to a final value given by the sum of the geometric progression in k:

Q (1 + к + к2 + etc) = Q x 1/1 — к (3.22)

This is illustrated in Fig 3.22.

In a reactor start-up, during the approach to cri­ticality, as the value of к increases towards unity the neutron population rises. If the control rods are withdrawn in stages, pausing between successive stages, the neutron population will settle at each stage at a value given by Equation (3.22). This behaviour is a clear indication to the operator that the reactor is subcritical. Because of the delayed neutrons and the time taken for the geometric progression in к to converge, it may take several minutes for the neutron population to settle.

At criticality, when к is exactly equal to unity, the neutron population will rise steadily according to the expression Q + Q + Q + … etc. Since Q is small, the rate of rise is slow.

At first sight this may appear to conflict with our understanding of criticality, which up to now has been that a chain reaction is just sustained and neutron population is constant. It must be emphasised, how­ever, that the presence of the neutron source is a special case in which the chain reaction set up by

the artificial introduction of a number of neutrons Q in each timestep is just sustained, and in each successive timestep a further Q neutrons is added.

In practice the linear rise at criticality is not im­portant. During a start-up it is usual to balance the reactor power at a low power of say 100 kW, which is several orders of magnitude above the source con­tribution, so the linear rise due to the source is not apparent. In this situation, criticality is indicated by the ability to sustain a steady power level.

In the foregoing analysis we supposed that a neu­tron source was suddenly introduced into the reactor core. In practice the neutron source is usually loaded into the reactor core during commissioning and re­mains there for the life of the reactor. Sb-124 has a haltlile of approximately 60 days, so the strength of the neutron source decays with time. Sb-123, one of the stable isotopes ot antimony, is an absorber of neutrons, changing to Sb-124 when it absorbs the neutron, therefore the neutron source is reactivated while the reactor is at power. Unless the reactor is shut down for a long period of time, it will there­fore not normally be necessary to remove the neu­tron source from the reactor core during the life of the reactor. Also, when the reactor has been operating at power for some time, the neutron source is sup­plemented by neutrons derived from some of the heavy nuclei formed in the fuel and from r, ,n) re­actions induced in various reactor materials by the у activity arising from the decay of fission products; this supplementary source activity also decays with time.

At some CEGB stations which during their life have suffered shutdowns of the order of 2 years after having operated for several years, special procedures were necessary to undertake the start-ups following the long shutdowns because of the low’ neutron source strength in the reactor cores, which meant that there was insufficient neutron flux in the cores to give an adequate reading on the start-up instrumentation. The detailed procedure varied from station to station, but basically the coarse rods were pulled in stages of about 50 mN each with a pause of a few minutes between stages to allow the neutron flux to reach equilibrium and to check if the flux was giving a meaningful reading on the start-up instrumentation; when the flux was giving an adequate reading the start-up could proceed normally. A similar procedure was adopted for the initial start-up of the AGRs, which were fitted with non-activated neutron sources and relied on spontaneous fission for the neutrons necessary for start-up.

For the purpose of describing the principles of the kinetic behaviour of reactor parameters in this sec­tion, a blower/circulator failure has been used as an example. Let us now consider controlled gas flow changes, for example, a reduction in gas flow to effect a reduction in reactor power for whatever rea­son. In most cases this can be carried out at a rate determined by the reactor control engineer, so many of the potential problems outlined in the preceding paragraphs should not occur. If a large reduction in power is intended, it will probably be necessary to estimate the magnitude of the increase in xenon worth which will result from the power reduction, to es­tablish whether there is sufficient positive reactivity available from pulling the control rods in order to maintain criticality. The reactivity available is often referred to as ‘xenon override capability*. Such an estimate is necessary on AGRs as well as magnox reactors, particularly if the reactor has a ‘backlog’ of refuelling so that the regulating rods are less deep­ly inserted into the core than normal. In this cal­culation, account can be taken of the duration of the load reduction, since if the operators are confi­dent that load will be restored before the xenon peak is reached, a larger load reduction can be tolerated because the load restoration will curtail the rise in xenon concentration.

Change in reactor gas inlet temperature with gas flow к has been assumed so far that reactor gas inlet tem­perature remains constant. In the case of a blower/ circulator failure, requiring the associated gas circuit to be taken quickly out of service, the reactor gas inlet temperature returned by the healthy gas cir­cuits will remain normal, while any deviation in the gas temperature from the failed blower/circulator will become irrelevant as the gas circuit is taken out of service. In considering gas flow changes on healthy gas circuits, for example, in a planned load reduction, on AGRs and at Oldbury and Wylfa (magnox reactors with once-through boilers), reactor gas inlet tempera­ture is held constant by auto control loops. On mag­nox reactors with drum boilers, however, no such auto control loops exist, and the thermodynamics of the boiler give a reduction in reactor gas inlet tempera­ture as gas flow is reduced. In addition there is a reduction in the temperature rise across the blower as blower power is reduced. The combined effect amounts to about 20°C for a 50^0 reduction in gas flow. Reactor gas inlet temperature can be controlled manually, see Section 5.5.4 of this chapter. A reduc­tion in reactor gas inlet temperature causes a reduc­tion in moderator temperature which is a disadvantage from the point of view of reactivity, so it is desirable for reactor gas inlet temperature to be held constant. This is also desirable from the reactor control engi­neer’s point of view because the fewer the number of variables which are changing the easier it is to control the reactor.

6.1.1 The requirement for monitoring

When a reactor unit is running at a steady appreciable power, absolute vigilance is still required although ac­tions are not occurring at the same rate as in the start-up or shutdown phase. Nevertheless, frequent monitoring of conditions and systems is necessary. The frequency of monitoring is preset in the operating in­structions and it is incumbent on the operator to carry out specific items of monitoring at stated times or frequency.

Two risks are present:

• The risk of commercial loss from either lost gen­eration or damage to plant and equipment.

* The risk of a nuclear accident from a rupture of the pressure containment accompanied by the failure of the nuclear fuel.

Although commercial loss is a serious event for CEGB financial considerations, it is not a danger to the public in that they can suffer injury to health, unless accompanied by a nuclear incident. Therefore those items of plant and equipment in which failure cannot result in a nuclear accident are monitored and protected in the same way as any commercial plant. Special considerations to protect the public are not necessary, thus a high proportion of plant and equipment on a nuclear power station is standard equipment similar to that in a conventional station.

The second risk of nuclear incident is the one that gives concern and is treated either to eliminate it or reduce it to negligible proportions (of the order of 1 in 106). Nuclear incidents that can affect the public in any serious way can only result from damage to the fuel whereby fission products are released into the gas stream and gas is lost from the reactor by way of a breach of the pressure circuit. With this effect in mind it is necessary to see how conditions can be monitored to give the operator advance warning or information for him to correct the situation.

The following list gives those items that need special consideration to ensure safe operation:

• Temperature of fuel, core and pressure vessel.

• Condition of fuel, to detect mechanical failure of the fuel cladding material.

• Loss of coolant flow due to circulator failure re­sulting in circulator run down.

• Loss of coolant due to failure in the pressure circuit.

• Run away nuclear power could result in damage to fuel by overheating.

These aspects are considered in the following para­graphs which address the systems that give informa­tion to the operator to enable him to manage his plant. It must be stressed that the probability of fail­ure leading to a nuclear incident Is very small, but it is essential that the operator is able to intercept any conditions which are not normal.

A general layout of the coolant control plant in a AGR is shown in Fig 1.41. The flow rate through the plant is achieved by utilising the pressure drop across the primary coolant circulators and is typically 50-100 t/h CO2, leaving the primary circuit at approximately 280°C. Minor variations occur from reactor to reactor dependent on the details of the plant duty required but normally this plant consists of:

• An inlet filter to remove activated metal oxide debris and prevent it from reaching the bypass plant in order to maintain a low activity level. The filter can be of the sintered stainless steel type with a typical efficiency of 100% for a 4 fim panicle.

Fic. 1.41 CAGR coolant control plant

• The flow is split and typically 10-20 t/h CO: is fed to the recombination unit at 280°C, Oxygen from either the electrolysis plant or the bottle sup­ply is also fed to the unit to reform carbon di­oxide. Typical oxygen flow rates are 5-15 kg/h aivine a 20-50^0 conversion of carbon monoxide across the bed. This both maintains the required carbon monoxide concentration in the main circuit and ensures that no oxygen, which can react rapidly with the araphite fuel sleeves and also more slowly with the graphite core, is fed back to the main coolant circuit.

The recombination unit catalyst is a 0.3-0.5% platinum on alumina pellet of 3.2 mm right cy­linders with a total bed weight of 0.8 t. The catalyst will remove up to l. O^o of its weight of sulphur which exists in the reactor coolant as carbonyl sul­phide at a typical concentration of 100 vpb, higher levels affecting steel oxidation and oxide spalling. The sources of sulphur are ingress of circulator oil, the graphite core and, where installed, the carbon shield. The adsorbed sulphur will slightly reduce the catalyst efficiency and this has to be allowed for in the recombination unit bed design.

• The recombination unit flow is recombined with the main bypass flow and then passes through one side of a regenerative heat exchanger which cools the gas to 120°C, the other side being the returned reactor gas. The bypass flow is then further cooled to 35°C by passage through a water-cooled heat exchanger.

• The gas is passed to the drier system which con­sists of two drier towers, one on duty and the other either being regenerated or on standby. The drier material is silica gel spheres of 3 mm diameter hav­ing a total bed weight of 5.5 t. The capacity of silica gel decreases with number of regenerations and is dependent on the temperature of adsorption, moisture concentration and regeneration tempera­ture. Typical design parameters are adsorption tem­perature 35°C, regeneration temperature 200°C when its capacity at 300 vpm H2O is 2-4%. The beds are operated on a 8-24 h cycle and can be changed either on a time basis or a moisture break­through basis. The beds are regenerated by taking a How of 10 t/h at 200°C from the regenerative heat exchanger with reverse flow through the bed, then through a water cooler and separator at 40°C to condense and remove the water from the circuit; linallv the flow recombines vvith the main flow and is passed through the drier bed in use.

• The required quantity of methane, typically 1-5 kg h is ted to the drier outlet flow from either the methanation plant or bottle supply.

• The gas is passed through the bypass plant outlet liber which may be of the porous stainless steel t>pe and mas comprise either a single stage or a dual coarse/fine filter. This filter will prevent any recombination unit or drier dusts being fed to the reactor and for a dual stage filter operates at effi­ciencies of 100% for 3 and 1 diameter particles respectively.

Mechanism of drier adsorption The breakthrough of a constant feed of adsorbate from a fixed bed is characteristic of the equilibrium relationship between adsorbate and sorbent and the mechanisms controlling mass transfer from the gas to the solid phase. The variation of the exit adsorbate concentration with time in a fixed bed at constant inlet concentration is affected by several factors Involving gas and solid properties and bed dimensions. These are all implicitly included in three main headings:

• The mass transfer step (or steps) which controls the rate of flow from the gas phase to the internal surfaces of the adsorbent. This may be gas film diffusion, intraparticle diffusion, surface adsorption or any combination of these.

• The equilibrium behaviour between gas and solid (the isotherm).

• The mass balance across the bed.

The adsorption of water by silica gel in the presence of high pressure carbon dioxide gives a linear isotherm and for such a system.

c 1 1 .

—- = — + — eaUx) sin a](x) x dx/x

Co 2 гг J о

where

Co is the bed inlet water concentration C is the bed outlet water concentration о і and аг are functions of the integration parameter x is the integration parameter

The input data required to solve this equation are:

• Gas properties — linear flow rate, density,

viscosity, Schmidt number (д/pD)

• Solid properties — particle density, average

radius, bulk density, bed depth.

• Gas/solid properties — particle Reynolds number,

adsorption isotherm gra­dient, intraparticle diffu­sion coefficient.

10.2 Graphite

Graphite is a pure crystalline form of carbon vvith a crystal density of 2.26 g/cm3. The properties of the
bulk graphite can be significantly varied, being de­pendent on the choice of raw materials and the manu­facturing parameters. The basic process is similar for all artificial graphites and starts with the conversion of a high molecular weight hydrocarbon, either a natural pitch or a residue of crude oil refining, to a coke by heating in the absence of air at 800°C — I000°C. The coke is then calcined, crushed and sieved to give the required particle sizes, typically 1-2 mm and below. The coke is mixed with hot pitch and either extruded or moulded to the size required. The brick at this stage is not dimensionally stable and is heated to 900°C-1000°C to coke the pitch and give dimensional stability. Volatiles are released during this process, leading to an extensive network of pores throughout the brick and to a relatively low bulk density. The brick may then be further impregnated with pitch and carbonised, several times, to increase the bulk density to the required value and each time affecting the distribution and size of the open and closed porosity. The material is graphitised by passing a current through a bed of the carbonised bricks surrounded by coke which heats the bricks to 2800°C — 3000°C. This process also leads to the volatilisation of most impurities. Purification can be assisted by the addition of either solid or gaseous halides to the graphitising furnace leading to the volatilisation of the more volatile halides, most importantly boron.

The properties of the three types of graphite used in the main cores of the magnox and AGR reactors are given in Table 1.16.

The BTRS is a subsystem of the CVCS and is shown diagrammatically in Fig 1.65. Its function in support of the RCS chemistry is to remove and store boron on anion resin at low temperature (10°C), and subse­quently release it on demand from the resin at higher temperature (60°C). By this means the required control of the boron concentration in the RCS is achieved in an economic way.

The boron capacity of the resin is proportional to the boron concentration in solution and inversely pro-

(2) Solids concentration determined by filtration through a filter having 0.45 micron pore size.

(3) Excluding boric acid.

portional to the water temperature. Low temperature operation is achieved by the use of chiller units and higher temperature operation by the use of modulated process heat.

It should be noted that although the BTRS was included in most PWR reference designs prior to 1980, it was not always operated. It can be argued that the operational advantages which led to the original con­ception of the BTRS, can be outweighed by a com­bination of safety related plant layout requirements, operational considerations and economics. It is possi­ble therefore that not all PWR designs will feature a BTRS type system, and may quite justifiably achieve an acceptable level of boron recovery by other means such as evaporation.

Fuel handling operations are dominated by the radio­logical safety of the operators and the public. Op­erations, both manual and automatic, must be se­quenced so that each step is monitored and correctly completed before the next step starts and other op­erations are locked out. This requires an extensive and complex control/interlock system with mechanical baulks and a multitude of microswitches showing equipment component position, together with mea­surement of parameters such as hoist loads. However, malfunctions must be expected. No malfunction must lead to a radiological hazard. Recovery must be pos­sible by interlock over-riding and handwinding when applicable.

Radiological shielding and pressurisation severely restrict the ability to see what is going on. Periscopes, TV cameras and shielded windows provide limited viewing for normal and recovery operations.