The graphite core of the reactor consists of machined graphite bricks which are locked together with either graphite kes or zirconium pins. The integrity of this structure must be maintained, since it must withstand
the loads imposed by thermal expansion and by di­mensional changes induced by irradiation. It must also withstand the static weight of the core itself and the associated fuel which together weigh some 1000-3500 t. The carbon dioxide (CCb) coolant of the magnox and

AGR reactors reacts with the core graphite, causing a «eight loss and a consequent loss of strength. It is the loss of strength rather than the loss of mod­erating effectiveness which necessitates the control of tzraphite oxidation.

The process of oxidation may be thermal or radio — Ivtic, the detailed chemistry of the processes being discussed in Chapter I. Since the thermal reaction between the graphite and coolant is insignificant below 600°C, this reaction does not predominate in CEGB reactors. Radiolytic oxidation occurs when CO: is decomposed by ionising radiation to gie rise to re­active oxidising species (positive ions resulting from absorption of energy by the CO:), Some of these ions are able to combine with the graphite and produce carbon monoxide (CO). The oxidising species also combine with the carbon monoxide to reform carbon dioxide. Since carbon monoxide is also formed by radiolysis of the coolant gas (CO:), its concentration is allowed to build up to inhibit the action of the oxi­dising species on the graphite. There is no further increase in inhibition at CO concentrations above 1.5 vol<?o, so that this level is the target maximum for the magnox range of reactors.

Graphite oxidation is a function of radiation in­tensity and gas pressure. The criteria of 1.5 voIVo CO (maximum) was acceptable for the early magnox re­actors. However, the rate of graphite oxidation in­creases through the magnox series mainly because the coolant pressures rise from 9 bar (Berkeley) to 27 bar (Wylfa). A detailed research programme identified hy­drogen as the most suitable inhibitor for the later stations. Ingress of water from boilers and oil from gas circulators give rise to low concentrations of hydrogenous compounds and the total hydrogen equi­valent from (H2O + H2 + CH4) is used for control purposes. It should be noted that at Hinkley Point A,~ the gas circulators have air seals and hydrogen may need to be injected to maintain the required levels. The use of hydrogen brings its own problem since high concentrations accelerate the oxidation of steels. Figure 3.65 illustrates the conflict of requirements with respect to graphite and steel oxidation, i. e., low hydrogen concentrations for steel oxidation control but high levels for control of the graphite reaction. The magnox system chemistry is illustrated in block diagram form by Fig 3.66.

Each reactor presents its own particular problems in coolant chemical control. Much depends on past operating conditions and coolant compositions. Con­trol of CO and hydrogen concentrations is exercised by periodic purges of the CO:, and in some cases a catalytic recombination unit is employed for CO control to combat the uneconomic use of CO: for purging. Gas driers are used for water removal but these are generally employed to remove water from the gas circuit following a prolonged shutdown for maintenance. Those reactors in which air is admitted to the coolant circuit at these times are provided with

‘O’v’.W a’ — ‘-1.

2T

dry air and the circuit is kept at a positive pressure. This is to reduce the quantity of vvater absorbed by the graphite and is primarily intended to reduce the dry-out time at the subsequent start-up of the reactor.

The AGR reactors, operating at higher gas pres­sures, temperatures and flux levels, present the graph­ite core with a more hostile environment than that of the magnox reactors, resulting in the need for a high degree of oxidation inhibition.

— Radiolytic oxidation takes place mainly within the pores of the graphite rather than on exposed surfaces. This is because the oxidising species have extremely short lifetimes and are quickly eliminated within a short distance of their point of production. This fact led to the development of reactor-grade graphites (pro­duced from Gilsonite deposits in Utah, USA), with a reduced pore volume compared with that used in the magnox reactors. The oxidation rate is also dependent on the pore diameter spectrum and geometry. The graphite should preferably contain a small number of large pores rather than a large number of small pores. It should be noted that oxidation increases the pore volume so that the rate of oxidation increases with time.

As in the case of the magnox reactors, carbon monoxide is an inhibitor of the oxidation rate but is of limited value (maximum factor of 2) and has the additional disadvantage of a tendency to form carbonaceous deposits, Further inhibition of the graph­ite corrosion reaction is obtained by providing a sa­crificial carbonaceous film on the surface of the graphite, derived from the radiolytic decomposition of methane. Unfortunately, this process gives rise to carbon deposits on the fuel pins, thus impairing heat transfer.

It has been established that methane is a powerful inhibitor (1000 vpm reduces reaction rate by a factor of 20). However, due to its radiolytic destruction in-core, technical and economic problems limit the maximum concentration that can be used in practice. For example, the use of high concentrations of meth­ane requires a large plant for the production of the gas, a large drier unit to remove the water resulting from the decomposition of the methane, and a re­combination unit to remove the CO so produced. Ob­viously, there is an economic limit to the size of such

equipment and this reflects the level to which the methane concentration can be raised (Fig 3.67).

It has been indicated that the production of the oxidising species takes place within the graphite pores. To gain access to the pores, methane has to diffuse into the graphite and, as a result, marked weight-loss profiles occur within a brick structure despite the provi­sion of methane access holes (Fig 3.68).

The relationship between the coolant composition and. graphite lifetime is shown in Fig 3.69. The term ’effective weight loss’ is a parameter which describes

the graphite weight loss in a brick since this is not uniform throughout. The highest effective weight loss that could be tolerated around the end of design life is presently considered to be about 20%. In Fig 3.69, the corrosion contours link gas compositions of equal effective weight loss and show that high methane concentrations prolong graphite life. Since plant economics preclude the adoption of certain composi­tions, these are bounded by the ‘plant limit’ line. Compositions that are prone to produce carbonaceous deposits are defined as those occurring above the predicted ‘deposition boundary’.

The initial gas composition selected for the CEGB’s lirst AGR commissioned was well below the deposition boundary — (1% СО/130 vpm CHj). Post-irradia­tion examination of the discharged fuel confirmed that carbon deposition was absent (some deposit was pre­sent but it was identified as being derived from lu­bricating oil). It will be noted from Fig 3.69 that the selected composition was far from ideal, in that a brick life of less than 20 years would be expected.

Subsequent to this initial trial, tests were carried out on the reactors at Hinkley Point В and Hunterston В with small changes to the gas compositions such that more inhibiting conditions were achieved. These changes advanced the expected brick life to about 28 years. However, there is a necessary delay between the completion of a trial and the examination of this fuel to assess deposition of carbon.

To overcome the delay in assessment and to pro­vide a more flexible method of detecting deposition, instrumental fuel stringers were developed. These units were provided with a number of thick-walled cans (1.8 mm instead of the normal 0.38 mm) with a hole drilled in the wall to accept a 0.5 mm thermocouple. The fuel pellets used in these cans were specially en­riched to compensate for the additional steel present in these particular cans. Thus the measured can wall temperature matched that of the standard fuel can. In addition, prqansion was made to measure gas inlet and outlet temperatures together with the gas-mass flow. The use of these stringers to detect deposition, requires that the temperature can be calculated taking into account all variables except deposition. The difference between the prediction and the as-measured can tem­perature is a measure of the deposition present.

The first trial was started in October 1982 with a coolant composition of 1.5% CO/300 vpm CHj, It was predicted that there would be a 9% reduction

in the heat transfer from the lowest element within a nine month period. By November 1983, no change had been detected. Subsequent post irradiation ex­amination (PIE) of the fuel confirmed that deposits present were within the range established before the test began. The composition tested was therefore judged to be non-depositing; Fig 3,69 shows that it is strongly inhibiting and provides some 30 full power years of life.

Tests in 1984 ’85 with coolants containing l. S^o CO/415 vpm СНд and 1.2°“o СО/350 vpm CH4 were later shown by PIE of fuels to have produced signi­ficant fuel pin deposition. Pending further evaluation of reactor and research rig evidence, CO and CH4 levels have both been reduced to avoid the rush of further deposition. The present coolant composition is ITo CO/230 vpm CH4, which will provide approxi­mately 25 years of full power life.

9.2 Steel oxidation lifetime The first of the CEGB’s magnox stations was com­missioned in 1962, the design being based on infor­mation available at that time. In 1968, the results of rig work and examination of specimens removed from some of the magnox reactors became available. These showed that steel oxidation rates of in-core compo­nents were unacceptably high at the (then) current operating conditions. The minimum gas outlet tem­perature was limited to 360°C on all the CEGB re­actors (except Berkeley which remained at 355°C) to ensure that the station’s economic lifetime would be achieved. Since that time the limited gas temperatures have been raised, thereby recovering some of the con­sequential loss of power output.