History Overview
The very first man-made nuclear reactor, CP1, had no coolant at all, the second, X10, which achieved criticality in November 1943 at Oak Ridge, was cooled by a gas (air). But today, most nuclear reactors are water cooled (PWR, BWR, HWR, and RBMK, in that order). Developments began in the area of Gas-Cooled Fast Reactors (GCFR) in the period from roughly 1960 until 1980. During that period, the GCFR concept was expected to increase the breeding gain, the thermal efficiency of a nuclear power plant, and alleviate some of the problems associated with liquid metal coolants. During this period, the GCFR concept was found to be more challenging than liquid-metal-cooled reactors, and none were ever constructed. In year 2000, the second era of development in GCFR began with Generation IV reactors. The new GCFR concepts focus primarily on sustainable nuclear power, with very efficient resource use, minimum waste, and a very strong focus on (passive) safety.
Natural uranium, graphite-moderated reactors were developed in the United States during World War II for the conversion of 238U to 239pu for military purposes. Following the war, this type of reactor formed the basis for the nuclear weapons programs of the United States and several other nations. It is not surprising, therefore, that natural uranium-fueled reactors became the starting point for the nuclear power industry, especially in nations such as Great Britain and France, which at the time lacked the facilities for producing the enriched uranium necessary to fuel reactors of the light-water type. Both of these countries have now constructed diffusion plants, however, and recent versions of both British and French reactors use enriched fuel.
The original plutonium-producing reactors in the United States had a once through, open-cycle, water coolant system, while the British-production reactors utilized a once-through air-cooling system. However, for the British and French power reactors, a closed-cycle gas-cooling system was adopted early on. This provides containment and control over radioactive nuclides produced in or absorbed by the coolant. In these reactors, the coolant gas is CO2. This gas is not a strong absorber of thermal neutrons and it does not become excessively radioactive. At the same time, CO2 is chemically stable below 540°C and does not react with either the moderator or fuel.
In a world progressively dominated by the water cooled reactors, mostly PWR and BWR, gas cooling remained alive in the high temperature reactor families, prismatic, and pebble bed HTR, associated with graphite moderation.
Still marginal, gas cooling is also present among the “Generation IV” concepts, through the very high temperature reactor system aimed at both electricity generation and hydrogen production and the GFR, gas cooled fast neutrons reactor.
Introduction
There are several classifications of nuclear reactors in operation and design. The major differences in design generally include reactor core layout, the fuel configuration the moderator material, the coolant type and types of control rods. With specific designs, each type of reactor also has particular operating characteristics. This document present the data related to design, characteristics and some technical aspects of Gas cooled reactors (GCR’s).
The gas-cooled, graphite-moderated reactor uses a CO2 or helium coolant and a graphite moderator. Its use is limited to the United Kingdom; it is sometimes known as the Magnox reactor. A larger second-generation version is the advanced gas-cooled, graphite-moderated reactor (AGCR).
Gas Cooling
Despite its name, the real task of a coolant is not to cool the fuel but to transport heat from the reactor core to the boilers to produce steam for electricity generation or to generate process heat. In that respect, gases exhibit interesting qualities.
First, because the density of a gas is variable, its operating temperature can be chosen independently of the operating pressure. Thus, a high gas temperature can be used, limited only by the core and circuit materials, to give good steam conditions from the boiler and thus good conversion of heat to electrical energy through the resulting high turbine efficiency. The optimum pressure can be selected separately on considerations of safety and of the economies of pumping power and pressure circuit costs.
Gas has certain intrinsic safety advantages. It can undergo no phase change as a result of rising temperature or falling pressure, and so there cannot be any discontinuity in cooling under fault conditions, and flows and temperature can be predicted more simply and with greater confidence.
Continuity of fuel cooling for on-load refueling is more easily achieved with a gas. In addition with a gas, there is no risk of a fuel-coolant interaction of the kind that in certain circumstances could result from the dispersion of melted fuel in a liquid coolant. Finally, a gas carries a relatively low burden of activated corrosion products, gives low radiation levels for maintenance round the circuit, requires small active effluent plants, and gives rise to only low radiation doses to the operators.
Offsetting these virtues is the combination of low density and low specific heat of gases. Even with fairly high pressures, this requires comparatively large temperature differences to transfer the heat between the fuel and gas, and between the gas and the boiler surface. As a result core ratings are low and core and boilers need to be large. It also requires large volume flows to transport the heat, and therefore large circulator sizes and powers.
The question of what gas to use is associated with the choice of moderator to slow the neutrons to thermal velocities. The first gas-cooled nuclear reactor above mentioned used graphite. The good moderating properties of graphite, combined with its low neutron capture cross section, have led to it being used almost universally for gas-cooled reactors. Only a handful of experimental or demonstration reactors have been built with heavy water as moderator with pressure tubes containing the fuel and gas coolant, like the French 70 MWe EL4 at Brennilis. Indeed, graphite moderation has become almost synonymous with gas-cooled reactors. If gas-cooled fast breeders are developed, this couple will have to divorce.
The choice of coolant gas is influenced mainly by the thermodynamic, nuclear and chemical properties, and by its cost and supply. Table 01 gives properties of some of the candidates at 300oC, a typical temperature of interest.
For good heat transfer and heat transport with low pumping power, a gas with high specific heat and high molecular weight, or density, is desirable. The coolant needs to have a low neutron absorption, to give good neutron economy and to avoid a rise in core reactivity if the reactor accidentally loses pressure. It also needs to be stable under irradiation, and preferably to have low neutron-induced radioactivity. Good chemical stability and low corrosion are obviously important. Of the many gases available the choice narrows quickly. Two gases stand out as candidates: carbon dioxide which is dense, cheap, but not chemically fully inert and helium which is inert, has a high specific heat, but is costly. These two coolants have led to parallel lines of development in the family of gas-cooled reactors, sharing much common technology but with differing characteristics. The carbon dioxide reactors are characterized by relatively low specific core ratings, moderate temperatures, and large size – typically the natural uranium plants of the UK and France and the advanced gas-cooled reactors (AGRs). The helium reactors aim for high ratings, small size and high temperature: the family of high-temperature reactors, or HTRs.
Natural Uranium Fueled Classical Reactors
Most “first generation” gas-cooled power plants were designed and built in the UK– the Magnox series – and in France – the NUGG. Both countries started their nuclear generation program in the early 1950s. Both had access to sufficient quantities of uranium ore, but no heavy water or enrichment facilities, and this severely limited the choices available. Graphite has many advantages as a moderator as it absorbs few neutrons enabling the use of natural uranium as a fuel. The graphite industry was also a mature one as the material had been used for a long time in the electrochemical and electrometallurgical industries.
To use natural uranium in a graphite moderated reactor, the fuel must be in its metallic state in order to achieve a high enough density of fissile material. It must also be renewed at regular intervals to minimize the number of sterile captures by the fission products. The MAGNOX and NUGG reactors used bars of uranium clad in a magnesium alloy. These were inserted into channels in a massive graphite pile through which carbon dioxide was circulated under pressure. These reactors were built using fairly primitive technology – that available in France immediately after the war – but the poor slowing-down power of graphite meant that the size of the plants had to be large in order to achieve significant power levels, and this in turn led to a high capital cost. Their sensitivity to the xenon effect made them very inflexible in operation, but the ability to unload the fuel without having to shut down the reactor made it possible to produce almost pure 239Pu for military applications by short irradiation.
Early “production” reactors, Wind scale in the UK and G1 in France were cooled by air, but even before the 1957 fire of the Wind scale reactor (due to the sudden release of the energy stored in the graphite by the “Wigner” effect during an attempt to anneal the pile), it was decided to turn to carbon dioxide as a coolant: This gas was readily available, cheap, and well known in industry. It has good heat transfer characteristics (for a gas) and good neutronic properties. It is also chemically compatible with the use of graphite as the moderator and with the cladding material and fuel used, provided certain precautions are observed. In addition to the series described below, two Magnox were exported by the British industry to Italy (Latina) and Japan (TokaiMura), while a “sister ship” of the French St Laurent units was built in Spain (Vandellos).
The Magnox Family Reactor
The Magnox reactor was the forerunner of the Advanced Gas Cooled Reactor (AGR). Several were built and entered commercial service in the United Kingdom. Their subsequent performance and low fuel cost made them very economical electrical power producers in the United Kingdom. As they evolved into a more complex but more compact design they approached the design of the AGR and in the final version really only differed from the AGR in the type of fuel. Hence the Magnox and AGR can be considered essentially the same type of reactors.
The first nuclear electricity on the western side of the iron curtain was generated by dual purpose (weapon-grade plutonium production and power generation) Magnox plants located at Calder Hall, a station inaugurated by HM Queen Elizabeth II in October 1956. Eight 60 MWe reactors were built almost simultaneously by the UKAEA: 4 units at Calder Hall and 4 at Chapel-cross, in Scotland. The Calder Hall design was simple and reliable but was not very efficient (22% thermal efficiency). The gas pressure was limited to 6.9 bar and the maximum outlet temperature was kept at 3450C. Refueling was carried out off-load at atmospheric pressure. All those units were shut down in 2003 and 2004, having operated over 45 years.
In 1955, Great Britain decided to embark on a significant program of nuclear power plants. The Magnox (magnesium non-oxidizing) alloy used to clad the uranium rods gave its name to the series. Nine commercial power stations with twin plants were eventually built in England, Scotland, and Wales, totaling 5,000 MWe (Table 02).
Table 2
The fuel was in form of metal rods, 28 or 29 mm diameter and between 48 and 128 cm length according to the model, clad with magnesium alloyed with a little beryllium and aluminum. The cladding is finned to improve the heat transfer to the gas. Gas outlet temperature was in the range 340-4100C and refueling was carried out on-load, to improve availability. That made it possible to unload immediately any failed fuel element without shutting the plant down.
The Magnox reached their peak efficiency, 33.6%, in Oldbury A (Fig.01), the general layout of which was to inspire the following AGR series. Once-through boilers were integrated around the core in the central cavity of a pre-stressed concrete vessel (Table03).
Volumic power was low and, consequently, capital costs were high but fuel costs were low enough to make the stations economically competitive during the 1970s and 1980s. Toward the end of the Magnox construction program, in order to reduce the corrosion of mild steel by carbon dioxide, it was decided to restrict the outlet temperature below 360oC.
Table 3
Natural Uranium Graphite Gas (NUGG) Reactor
Nine NUGG reactors were built in France. The first three reactors, at Marcoule, were used almost exclusively for the production of plutonium. The electrical power generation program began with the successive commissioning of Chinon 1 (1957), Chinon 2 (1958), and Chinon 3 (1961), with power capabilities of 70,200 and 420 MWe net respectively. There was no question of waiting for these reactors to go critical, even less of waiting for the first operational results, before starting work on the next design. These three reactors were prototypes, and each very different from the others. The next reactors were built at Saint Laurent des Eaux (1963 and 1966) and Bugey (1965). The Fifth Plan (1966–1970) included plans to build a total capacity of 2,500MWe of NUGG reactors. The construction of a new unit at Fessenheim began in 1967, but was abandoned at the end of 1968. By that time, water-cooled reactors had become the favored option.
The characteristics of NUGG reactors are listed below, using Saint Laurent 1 as an example
(see Fig. 2 and Table 4).
The low specific power of the reactor meant that the core had to be very large. This core was enclosed in a vessel that also contained the coolant circuit and its heat exchanger. The vessel was a prestressed concrete structure, 33m in diameter and 48m high. The internal face of the vessel was lined with steel, 25 mm thick, in order to prevent any leakage of the CO2 under a pressure of 29 bar.
The graphite pile in the reactor was in the form of a vertical cylinder 10.2m high and 15.7m in diameter. It consisted of a network of columns locked together by mortise and tenon joints. The pile weighed no less than 2,680 tons. The four CO2-steam heat exchangers were single tube cross circulation types. The water inlet was in the lower section, while the hot CO2 entered the upper section. The total CO2 flow rate was 8.6 tons/s, and the steam flow rate was 0.6 tons/s. The fuel elements were replaced while the reactor was in operation, at a rate of around 2–3 channels per day, requiring the use of a sophisticated handling system.
A system for detecting the presence of fission gasses in the coolant was used to detect and locate any breaks in the claddings.
The fuel elements used in the NUGG reactors were developed over time. In the latest versions, each element consisted of a metal tube of uranium alloyed with 0.07% aluminum and 0.03%iron, surrounding a graphite core. The borderline neutron balance of the NUGG reactors resulted in a fairly low fuel burn up rate of 6.5 GWd/tons. The maximum operating temperature of the reactor was determined by the maximum permissible temperature of the uranium. This was set at 650oC at the internal surface of the tube.
The last NUGG plant, 540 MWe Bugey 1was shut down in 1994.
Table 4
Design Evolution
A significant design constraint in the Magnox reactors was the temperature at which the fuel could operate. Although pure uranium metal melts at 1130°C, it undergoes an α-β phase transition at 661°C. Associated with this transition is a volume change of about 1%. Any thermal cycling through this temperature thus leads to surface deformation and cavity formation. This limits the practical operating peak fuel temperature to not more than about 660°C. Furthermore the magnesium alloy, Magnox, has a low melting temperature of about 650°C so the cladding is limited to a temperature of not more than this value. These limitations in turn limited the maximum reactor coolant outlet temperature to about 400°C which limited the maximum steam temperature. Design refinements over the years and the use of various alloying materials for both fuel and cladding allowed coolant and steam temperatures to rise slightly with the result that capacity and efficiency increased with more advanced as shown in table 05.
Another significant parameter affecting the design and performance of Magnox reactors was the pressure of the carbon dioxide coolant. An increased pressure of the gas results in increased density and an increase in the rate of heat removal from the reactor core. Most of the early Magnox reactors had spherical steel pressure vessels surrounding the core and external coolant ducts leading to separate boilers as shown in Figure 03. This limited the pressure of the reactor coolant since the pressure vessel had to be large enough to accommodate the reactor core but its thickness not so great as to create manufacturing and erection difficulties. The first Magnox reactors had carbon dioxide pressures of less than 1 MPa but this was gradually increased to nearly 2 MPa as the design evolved as shown in Table 06. To go beyond this value required entirely new concept which was the development of the prestressed concrete pressure vessel.
The prestressed concrete pressure vessel as shown in Figure 04 was adopted for the last two Magnox plants and for the next generation of advanced gas cooled reactors. With this design the steam generators are located adjacent to and around the periphery of the core of the reactor. The prestressed concrete pressure vessel surrounds both the reactor core and boilers and serves to contain the reactor coolant under pressure and to provide the necessary biological shielding for the rest of the plant. The concrete being weak in tension is maintained in compression by steel tendons located in helical fashion around the circumferential shell and in a semi-radial manner across the top and bottom slabs. Compression in the concrete is obtained by post-tensioning the steel tendons separately after construction. The stress in the tendons can be monitored and they can be retensioned if necessary. The pressure that such a vessel call withstand is determined by the mesh of steel tendons thus allowing higher internal gas pressures than with a simple steel shell. Carbon dioxide pressures for the last two Magnox reactors are well above 2 MPa and for the next generation AGRs around 4 MPa.
A further constraint imposed by limited fuel temperatures and hence relatively low gas outlet temperatures is that of steam generation. For good efficiency in the steam cycle, feed water heating up to nearly saturated conditions is desirable. Most heat from the gas should be for evaporation and superheating. To match the gas conditions this requires a relatively low steam pressure which in turn is detrimental to cycle efficiency. However by adopting a dual pressure steam cycle, with an additional high pressure loop, the steam conditions can be made to better match the gas conditions and improve the thermodynamic efficiency. This naturally increases the complexity of the cycle. However with increased gas temperatures the single cycle could be used on the last Magnox plant and on the subsequent AGRS.
Advance Gas Cooled Reactor
The limitations inherent to the use of natural uranium were recognized from the start. In the mid-1950s, the British started studies of an improved design, based on low enriched uranium oxide fuel, manufactured in clusters of small diameter pins, with stainless steel cladding. This design was called AGR (Fig 05). The AGR is a direct descendent of the MAGNOX reactor and has only been built in Great Britain. After a small demonstration 30 MWe reactor commissioned at Windscale in 1962, a commercial AGR program of seven twin 600 MWe units was started (Table 07). The power density is four times that of a MAGNOX reactor, and the volume of the heat exchangers is smaller. The chemical compatibility of U02 and CO2 and refractory nature of the oxide make operation at higher temperatures possible. The coolant is at 650°C on leaving the core, giving the AGR an excellent electrical efficiency (42%). The first reactors suffered from a number of problems, partly due to failures in industrial organization, and partly due to a failure to control corrosion. Methane was added to the coolant gas in order to reduce radiolytic corrosion by CO and the oxidizing free radicals formed by the irradiation of the CO2. Controlling the concentration of this gas proved to be difficult.
This first generation of gas-cooled reactors has an excellent operating record, generating electrical power continuously with no major accidents. However, these old NUGG MAGNOX, and AGR designs are now obsolete for economic reasons. Graphite moderated gas cooled reactor technology has gradually been abandoned in France, Italy, Spain, and Japan, and only accounted for 4% of worldwide nuclear capacity in 2008. The British AGR and MAGNOX reactors are the only types still in operation. All of them should be decommissioned by 2020.
Figure 5: Cross-section of an advanced gas-cooled reactor (AGR) with single cavity vessel (Marshal W 1981)
Designing features of AGR (Advance Gas Cooled Reactor)
Figure 06 shows the main design Features of AGR;
Mechanical Design
In a typical AGR system, the reactor core, boilers and gas circulators are housed in a single cavity, pre-stressed concrete pressure vessel. The reactor moderator is a sixteen sided stack of graphite bricks, it is designed to act as a moderator and to provide individual channels for fuel assemblies, control devices and coolant flow (Figure 06 and Table 07).
The graphite is covered by an upper neutron shield of graphite and steel bricks and mounted on a lower neutron shield of graphite bricks that rests on steel plates. Radial shielding is in the form of steel rods located in two outer rings of graphite bricks. The graphite structure is maintained in position by a steel restraint tank that surrounds the graphite and is supported on a system of steel plates.
Figure 6: Main components of AGR
The shielding reduces radiation levels outside the core, so that when the reactor is shut down and depressurized, access to the boilers is possible.
There are two main effects of irradiation of the graphite moderator. One is dimensional change and the other is radiolytic oxidation by the carbon dioxide coolant. There will also be a significant change in the thermal conductivity, which decreases with an increasing temperature. Because of the relatively high temperatures in the core, there will be little or no stored energy in the graphite (Wigner energy).
The dimensional change due to the anisotropic properties of graphite is reduced by manufacturing the graphite bricks by use of moulding rather than extrusion.
When CO2 is radiolysed it breaks down giving CO and a very reactive chemical species which behaves like an oxygen atom
CO2 + CO → O (radiolytic)
Most of these species recombine in the gas phase
“O” + CO → CO2
However, some of them will escape recombination in this way - the mean distance that the active species can trave1 before undergoing reaction is slightly greater than the mean pore diameter of the graphite - and will reach the graphite surface where they will react:
“O” + C → C(0) (graphite surface reaction)
Where C (0) is a surface oxide. This will subsequently break loose to give gaseous CO. The consequences of the process are a weight loss of the graphite. However, only that gas contained in the pores of the graphite takes part in the reaction, the bulk of the gas in the reactor circuit is not involved.
The criterion adopted for the maximum permissible mean weight loss of graphite has been set to 5 % over a 30 years lifetime.
Radiolytic oxidation is inhibited by adding methane to the coolant. However, methane in high concentrations can lead to carbon depositions, in particular on the fuel assembly surfaces. Therefore, a compromise between protection of the moderator and deposition on the fuel must be made by a careful choice of the concentration of the inhibitor.
The core and the shield are completely enclosed by steel envelope called the gas baffle, the main function of which is to produce a downward flow of coolant gas (re-entrant flow) through paths in the graphite moderator to cool the graphite bricks and to separate the hot from the cold gas (Figure 07).
Figure 7: Gas baffle with gas flow paths
The gas baffle has three main sections - the dome, the cylinder and the skirt. In the doom there are a number of penetrations - one for each of the fuel channels in the graphite moderator. Between the penetrations and the tops of the channels, system of guide tubes provides the paths for the fuel assemblies and the Control rods. The skirt forms the lower part of the gas baffle cylinder.
The core and the radiation shields are supported on a structure called the diagrid, which itself forms an integral part of the gas baffle. This diagrid is designed to carry the weight of the reactor core and to accommodate the thermal movements which arise from coolant temperature variations during normal operating and in the case of incidents.
