The reliability and integrity claimed from the pro­tection systems to support the safety case have been based on demonstrable high quality equipment designs, good techniques and good practice supplemented by comprehensive testing. This also applies to the soft­ware that forms an integral part of the PPS.

Rigid methods of software preparation have been established which result in very comprehensive docu­mentation at all stages. This enables every module to be assessed by a third party against the specification requirements and for construction error. Comprehen — soe verification of software documents and validation testing to pre-defined test specifications is performed. In addition, an independent design assessment is car­ried out on the completed system.

The fail safe design of hardware requires the selec­tion of components to give a predictable output on failure. The software includes self-test features for hardware tault detection to produce defined actions In an equivalent manner.

The overall system software follows the principle that a computer failure is detectable by the next com­

puter in the operational chain and by cross checking the data output by readback to the transmitting com­puter. The software also includes many of the estab­lished self-diagnostic and continuous error checking techniques in general use in the industry,

In AGRs, the burn-out of fission cross-section results in a reduction of channel rating with burn up. There are two main components to this reduction, since the rating is defined by R = ]£fd>(E)$E and both Sr and <5 change with irradiation. Ef decreases {as shown in Section 2.2 of this chapter) due to the burn-out of U-235, and the flux d>(E) will reduce because the number of neutrons produced in the re­gion decreases and the neutron absorption cross-section increases as a result of the build-up of fission pro­ducts. In AGRs the channels are more widely spaced than in magnox and hence flux levels are not so dominated by surrounding channels.

Typical ‘age factor’ variations are shown in Fig 3.9, for fuel designed for 18 GVVd/t and 24 GWd/t mean

Fig. 3.9 Typical ‘age factor’ ariations for fuel
designed for 18 GWd/t and 24 GWd/t mean
channel discharge

local can temperatures in AGRs and therefore need to be taken into account in the assessment of can temperature for satisfying temperature operating lim­its. The burn-out of ring-to-ring rating ratio is illus­trated in Fig 3.10 for outer zone feed fuel. Burn-out effects are relatively smaller for the lower enriched inner zone and initial charge fuel; this is partly as a result of the smaller fine structure but also because the build-up of Pu-239 is relatively more important in lower enriched fuel.

Axial flux fine structure is a significant contributor to the peak rating and therefore affects can tempera­ture and fission gas release predictions. At start of life the rating is higher than the pin mean by about 20% in the end pellet and 10% in the next pellet. It has been observed from gamma scanning of irra­diated fuel that the peak in rating decreases signifi­cantly. A fit to measurements is shown on Fig 3.11, where the ratio of pellet mean to pin mean rating is given as a function of burn-up for end pellet and next pellet in the inner ring of 2.5% enriched fuel. Similar variations occur in the middle and outer rings.

Another localised rating factor is of significance. This is the cross-pin variation. The thermal flux dip into an AGR fuel cluster also generates a cross-pin flux tilt. This has an effect on cross-pin temperature distribution which is of concern in considerations not only of peak can temperature, but also of pin bowing

Мі; я. lu AGR ringno-rine raiing ratios

and clad straining due to differential expansion be­tween fuel and clad. The cross-pin rating tilt tends to burn-out since the burn-up is faster on the high flux side, but this is counteracted to some extent by enhanced Pu production on the outer facing surface. This is most pronounced in the outer fuel ring where the outer facing surface ‘sees’ directly the majority of the neutrons arriving from the channel wall. The resulting rating tilts as calculated by the WIMSE code are shown in Fig 3.12.

Before start-up commences, pre-start checks are car­ried out to ensure that the necessary plant and equip­ment are available so that the start-up can safely pro­ceed. In addition to these the plant is prepared in a number of other ways. For example, if the reactor has been depressurised for access to the gas side dur­ing the shutdown, the reactor gas circuit may contain some air which is displaced by purging with fresh CO2. This is generally done at low reactor gas pres­sure, say 2 bar, then the required purity is achieved with the minimum consumption of CO2. The Operat­ing Rules state a minimum gas purity which must be achieved for operation at power, but operational con­venience and good operating practice often require a much higher gas purity than the Operating Rules’ minimum requirement.

In order to ensure that adequate reactor gas flow is available to remove reactor heat, then not only must the required number of boilers and gas circu­lators be in service, but also there must be adequate reactor gas pressure. On the steel pressure vessel sta­tions, however, the pressurisation of the reactor gas circuit is limited by the pressure vessel metal tempera­ture, i. e., to avoid brittle fracture the pressure vessel must be adequately warm before the reactor gas pres­sure is raised. A limit of maximum gas pressure against vessel minimum metal temperature is given in the Operating Rules, a typical curve is shown in Fig 3.24. Pressure vessel metal temperature is raised by heat from the gas.

The heat to raise reactor gas temperature may be supplied in a number of ways. Many of the mag — nox drum boilers have a facility to accept warming steam into the boiler drum from the running reac­tor. A common method, particularly on reactors with once-through boilers where the warming steam method

Fie. 3.24 Limn of maximum gas pressure against
vessel minimum metal temperature
Operating Rules on magnox stations with steel pressure
essels specify the maximum permissible gas pressure
for a given vessel minimum metal temperature.

The limit shown is taken from the Hinkley Point A
Operating Rules.

is not possible, is to run the gas circulators, because most of the energy input to a gas circulator appears as heat in the gas.

After a prolonged shutdown of any reactor it may be necessary to apply heat to raise reactor gas inlet temperature to the minimum value required for start­up. A minimum value is specified for several rea­sons. First, it may be a requirement of the plant, for example, boiler outlet components or reactor core support structure. Second, it is to ensure that when the reactor is operating at power there is a balance between the accumulation of Wigner energy in the graphite due to neutron bombardment, and its an­nealing out so that the level of stored energy is ac­ceptably low {graphite temperatures of greater than about 160°C will ensure adequate annealing). Third, it is to ensure that the reactor is operated within the range of conditions which have been studied in fault >tudies and for which adequate protection has there­fore been demonstrated.

There may be limitations on the rate at w’hich re­actor gas inlet temperature can be raised, for example, due to stresses set up by temperature differentials in below — core components or the lower layers of graphite, also on some concrete pressure vessels there are limitations on rate of change of concrete tempera­ture, At Hartlepool and Heysham I this latter point, combined with the need to maintain a trickle feed on і be boilers w hich remov es some of the heat put in b> the gas circulators, constitute a severe constraint on the time taken to reach the minimum reactor gas inlet temperature for start-up.

6.4.1 Control

Over the upper part of the reactor power range, some of the parameters are under automatic control. When outside the range of auto control, the plant is con­trolled manually by the operator who uses the indi­cations to exercise control when the auto control is out of range or has failed.

It should be noted that not all CEGB nuclear stations are provided with automatic control, but it is provided on all the AGRs.

Usually power assessments are carried out and re­corded every eight hours. This enables the operator to look at reactor power historically and at the same time gives an indication of the relative accuracy of his reactor desk instrumentation. In addition it gives an accurate indication of normal reactor power so that a true estimate of xenon can be made if shutdown occurs.

6.2 Shutdown of a reactor system

A reactor shutdown may occur under these principal situations:

* A reactor may trip due either to fault conditions or to conditions arising that require the reactor to be tripped.

* Controlled shutdown where there is no urgency to complete the sequence and prior consideration can be given to each step.

* Emergency shutdown where it is essential to close the reactor down but does not warrant the manual trip button being operated to give immediate shut­down.

Where possible the controlled shutdown is preferred for two reasons, firstly, it is not good practice to thermally cycle the cores and structures of reactors since this accelerates wear and tear, and secondly, a controlled shutdown gives the operator more time to prepare and think out his course of action thus avert­ing any mistakes which may have an adverse effect on plant and machinery.

4.11 Fuels

4.11.1 Introduction

The following sections, which deal with nuclear fuel, are concerned almost entirely with those commercial reactor systems which are currently operating, or are in the course of design and construction in the UK. This gives a natural division between the indigenous gas cooled systems (magnox and the commercial ad­vanced gas cooled reactor CAGR) and the USA devel­oped pressurised water reactor (PWR) which evolved from the small reactors originally designed for opera­tion in nuclear submarines.

Harris and Duckworth (1982) [1] have shown how the British magnox reactors grew from the weapons programme and how the choice of natural uranium fuel was dictated largely by circumstances, rather than a systematic search for the best solution to an engi­neering problem. Nonetheless, natural uranium metal has been found to be a highly satisfactory nuclear fuel with several advantages over other types. It is the densest form of uranium and, with a graphite mod­erator, can be used without enrichment. It is readily available in a fairly pure form and it has a high co­efficient of thermal conductivity. This last factor is important since a major problem with metallic ura­nium is that it cannot be operated at temperatures much above 660°C; the high thermal conductivity al­lows this constraint to be met fairly easily whilst still producing gas outlet temperatures which allow rea­sonable turbine thermal efficiency. Provided that there is a sufficient concentration of fissile atoms to achieve criticality, the thermal conductivity of the fuel is pro­bably the most important factor in deciding the fuel shape. This is because the thermal conductivity has a direct bearing upon the temperatures which, as al­ready observed for uranium metal, generally sets the operating limits for the fuel.

In the case of uranium dioxide, the coefficient of thermal conductivity is low (between 4 and 17 times lower than for pure uranium) so that, although the allowable operating temperature for the oxide is much higher (melting temperature 2800°C) than for the metal, the fuel diameter has to be considerably smaller (approximately halved) in order that it can be ade­quately cooled. This separation of the fuel to allow cooling, together with the lower density of the oxide, usually dictates the need for enrichment of the U-235 (in the CAGR the use of stainless steel clad makes this essential). The balance of the fuel comprises the fertile isotope U-238 which can be converted by neu­tron capture and beta decay to (fissionable) Pu 239.

By the use of enriched fuel, smaller pins and more efficient cooling, the rate of heat generation per unit mass of fuel, or rating has been progressively in­creased: magnox reactors and CAGRs have peak ele­ment ratings of about 5 and 20 MW/t respectively; in PWR the maximum fuel rod rating is about 60 MW/t. The increase in rating between magnox and CAGR is largely attributable to the higher fuel tem­peratures possible with oxide fuel. The difference be­tween CAGR and PWR is mainly the result of the increased effectiveness of pressurised water cooling (compared to gas cooling) which allows these ratings to be reached whilst maintaining the fuel at accept­able temperatures.

In this section frequent reference will be made to the fuel burn-up. This can be expressed in three ways (e. g,, see Olander, 1976 [2]); firstly, the fission density, or the total number of fissions/unit volume; secondly, as a fractional burn-up, or the total number of fissions divided by the initial number of heavy metal (not necessarily fissile) atoms; or, thirdly, as the thermal energy released by one tonne of heavy metal atoms (noting that 210 MeV/fission is equivalent to 0.95 MWd/g fissioned) — this is the measure adopted here. For conversion purposes, 1% fractional heavy metal burn-up is equal to 8.6 GWd/t. It is important to remember that the fuel rating, and hence the burn — up, will vary from point to point in the reactor. Thus in CAGRs, for example, the peak stringer burn — up will be less than the peak element burn-up, which will be less than the peak pin burn-up and the peak point burn-up.

In general, the lifetime of the fuel will be deter­mined by its endurance which we may loosely define as the maximum burn-up attainable by the fuel be­fore significant numbers of failures begin to occur; this is the main subject of this section. In addition to this, however, there is also a limit on the maximum achievable burn-up which is dictated by the reactor physics; this is the point at which the reactor runs out of reactivity. In a reactor using enriched fuel the burn-up limit can be increased, at least in theory, by increasing the fuel enrichment, although in some cases it may also be necessary to use burnable poisons to maintain a more uniform reactivity throughout the life of the fuel. In reactors which are fuelled with natural uranium, however, this option is unavailable and in the magnox reactors, where the endurance limit has been steadily increased over the years, the reacti­vity limit is now being approached.

High purity water is used in nuclear systems as a primary reactor coolant (pressurised water reactor), for steam raising in a secondary circuit (advanced gas cooled reactor), or fulfilling both roles (boiling water reactor). This section deals with the applications where the water passes through the core but does not con­stitute the steam raising circuit. It is specific therefore to the PWR, where the reactor coolant is high tem­perature water held in a single phase at a pressure above its saturation vapour pressure, for the purposes of neutron moderation and heat transfer.

Typical reactor coolant conditions for steady state full power operation of a 1200 MW(e) PWR are shown in Table 1.17, and the main thermodynamic and physical properties of water at 155 x 105 N/m2 (155 bar) taken from Grigull et at (1984) [25] over the tem­perature range of interest are summarised in Table 1.18. On exit from the reactor there is, in principle, an approximately 20°C margin to boiling and the system therefore is single phase, although in some de­signs it is expected that a degree of boiling will occur at the top of the fuel elements. Maintenance of sys­tem pressure and adjustment of the coolant inven­tory are achieved by the pressuriser (Section 9.2.1 of this chapter) which contains coolant under SVP con­ditions of 155 x 105 N/m2 and 344.8°C in an atmos­phere of hydrogen.

It should be noted that the 240 m3 coolant in a typical PWR undergoes a significant specific volume change over the temperature range of 25°C at atmos­pheric pressure (1 x 105 N/m2) to 30O°C at operating pressure (155 x 105 N/m2), as can be seen from the specific volume data in Table 1.18.

1.4.1 Pressure circuit — primary circuit

Function

The function of the reactor pressure circuit is to contain the coolant gas which transfers the core­generated heat to the boilers. An increase in gas pressure increases the rate of heat removal from the core and decreases the circulator power. As successive magnox stations were developed, gas pressures were increased.

At the stations with steel pressure vessels, the de­sign limited the gas pressure which could be used for a given size of core and hence reactor output. This was the primary incentive for adopting the pre-stressed concrete pressure vessel at Oldbury and Wylfa (Table 2.1).

Standards and design

Safety considerations require that the components of the primary coolant circuit have the highest integrity. When the first stations were designed, it was generally agreed that the existing standards for the design, con­struction, inspection and testing of these components were not adequate. Supplementary requirements to the existing standards were agreed with the Independ­ent Inspecting Authority and specified for primary coolant circuit components. These supplementary re­quirements were mainly concerned with the detailed analysis of the design and the quality of the fabrica­tion and inspection of the reactor components. They now form part of the relevant British Standards.

The Nuclear Installation Inspectorate (Nil), has a condition in the Station Site Licence which requires the CEGB to appoint an Independent Inspecting Authority.

Steel pressure vessel — reactor pressure circuit layout The reactor pressure vessel containing the core is con­nected to the boilers by ducts. Typical arrangements of the pressure circuits are shown in Fig 2.8, the number of boilers per reactor varying from station to station.

The location of the boilers relative to the reactor is governed by a number of considerations which include:

• The need for adequate shielding between the core and the gas circuit components.

• Gas distribution to and from the core.

• Economics (capital cost and gas circuit pressure drop).

• The requirement to have a good natural circulation round the gas circuits. This influences the relative heights of the boilers and the core.

Each gas circuit consists of inlet and outlet ducts connecting the reactor to the boilers; gas valves to isolate the reactor from the boilers; bellows units to cater for component thermal expansion; a gas circu­lator to circulate the coolant gas round the circuit and a bypass circuit which connects the outlet of the gas circulator to the boiler, i. e., bypassing the reactor.

Boron trifluoride type

The boron trifluoride (BF3) proportional counter de­tects neutrons by the n-a boron-10 reaction and con­sists essentially of a fine wire concentric with, and insulated from, a thin copper or aluminium tube to form a vessel filled with BF3, as shown in Fig 2.47.

The electrons produced in the primary ionised tracks are attracted to the central anode wire and, with a sufficiently high positive potential (1.5 to 2 kV), sec­ondary ionising collisions occur and the final collected charge is thus amplified by a gas multiplication pro­cess. The BF3~gas is free from electron capturing impurities and the counter body is thoroughly out — gassed to prevent subsequent contamination of the gas and deterioration of the counting characteristics. Good vacuum characteristics are obtain. J by using oxygen-free copper and fluxless brazing techniques for all joints.

Modern counters also use aluminium extensively and this gives a better performance than copper, par­ticularly for life in high neutron fluxes.

For a 12EB40 counter filled with enriched boron triflouride to 400 mm Hg, the slow neutron sensi­tivity is about 3.5 counts/s/n/cm2/s. This type of counter will work satisfactorily in a у flux of about 2 Gy/h and the maximum counting rate possible is about 5 x 104 counts/s because of pulse length. Proportional counters are about 100 times more sen­sitive to slow neutrons than fission counters of simi­lar size, but unfortunately their 7 sensitivity is also considerably greater and it is not possible to discri­minate between 7 and neutron pulses of similar amplitude.

With a gas multiplication factor of 40, the life of a BF3 counter is about 1010 counts at below 100°C whilst above 150°C counters deteriorate rapidly. Counters are not operated in neutron fluxes above about 2 x 107n/cm2/s to avoid difficulty in discri­minating between neutron pulses due to space charge effects.

The BFi proportional counter is used extensively in the thermal columns of magnox stations with arrangements for withdrawal during power raise and insertion during the shutdown phase, as discussed in Chapter 3.

Fission type

By the use of the fission process, detectors can be made that operate at higher temperatures and neu­tron fluxes than the BF3 type without suffering from

pulse pile up.

A typical uranium-235 fission counter consists of a stainless-steel container with cathode liner concen­tric with, and insulated from, the cylindrical anode. A thin layer of fissionable material is deposited on

the electrode surface.

The fission counter detects neutrons by the U-235-n reaction; the resulting fission fragments cause ionisa­tion of the filling gas. The coating of uranium is limited to about 1 mg/cm2 to limit energy lost by fission fragments emerging from the coating. The neutron-induced fission pulses, relative to the 7 pulses in this type of counter, are much larger than in the proportional counter; they are also much shorter in duration, and these two effects enable the fission counter to operate satisfactorily in 7 fluxes up to 103 Gy/h. The sensitivity of the P7 counter, shown in Fig 2.48, is in the range 0.01 to 0.1 counts/s/n/cm2/s with rates up to 2 x 105 pulses/s with 10% counting losses.

The operating potential required is between 200 and 400 V. Since inert gases are used for filling and there_is no gas multiplication, fission counters are less sensitive to impurities and may be operated at temperatures of up to 550°C and with containment. up to 40 bar The life of the counter is 1019 nvt and it will operate in a neutron flux of about 1011/ cm2/s before j3/pulse ‘pile-up’ becomes serious, thus limiting the use of the detector at low powers after

irradiation. The burn-up at 10n n/cm-/s is 0.2ro per annum. Further details of fission counters type P7 are given in Table 2.5.

Another chamber, P8, is available specially devel­oped for Campbell channels, described in Section 5.2.12 of this chapter.

The reactors for the AGR stations, like those of the later magnox stations, are contained in concrete pres­sure vessels. The gas pressures and temperatures are higher and, with the exception of Hartlepool and Hey­sham /, the boilers are situated in an annulus round the core.

At Hartlepool and Heysham 1 the boilers are cylin­drical in shape and are located in the walls of the vessel. Horizontal ducts in the vessel walls connect the boilers to the main reactor vault. With this de­sign of station, the removal and replacement of boiler
units, although difficult, is easier than with other de­signs of AGRs.

Figure 2.80 shows the layout of reactor components within the concrete pressure vessel at Heysham 2, and the gas flow round the circuit.

In the AGR reactor design, the core graphite is cooled by gas at reactor gas inlet temperature which flows downwards through annuli round the fuel strin­gers. This gas then joins remaining gas and flows ver­tically upward through the fuel channels to transport the heat from the fuel to the boilers. To provide passages for the direction of the reactor coolant from the gas circulators into and out of the graphite core and to the boilers, large steel structures are necessary to separate the gas at reactor inlet pressure from the gas at reactor outlet pressure. These structures are called pressure cylinders and domes at Dungeness В, gas baffles at Hinkley Point B, Hunterston B, Hey­sham 2/Torness and hot boxes at Hartlepool and Heysham Ї,