Materials help extend UK nuclear site life
Analysis of the materials used in the UK’s ageing nuclear reactors shows there may be life in them yet.
Nuclear energy contributed 19.5% of the UK’s on-grid electricity capacity in 2018, according to the UK Department for Business, Energy & Industrial Strategy. But amid an ongoing national dialogue over nuclear newbuilds – which focuses on the two reactors under construction at Hinkley Point C in Somerset – the nation’s existing fleet of nuclear power reactors continues to age.
Keeping existing plants running for as long as possible while maintaining the most rigorous standards of safety presents profound challenges, which have been taken up by universities and research centres around the UK. Many of these, including the universities of Bristol, Plymouth and Southampton, have joined together to form the South West Nuclear Hub, to collectively find solutions to safely extend nuclear reactor lifespans.
Unique reactor design
Understanding each reactor’s design and materials helps ascertain what can be done. The UK’s nuclear energy industry has almost exclusively operated reactors cooled by circulating carbon dioxide gas and using graphite as a neutron moderator. The world’s first civil nuclear power station opened at Calder Hall, Cumbria, in 1956. It contrasts with other nations that use light or heavy water-cooled power reactors (WCR) – although a variety of different water-cooled designs exist worldwide. The design of UK power reactors has led to unique materials challenges at construction stage and during operation.
Magnox, the first generation of UK power reactors, were used to produce plutonium for nuclear weapons as well as electricity. In 1995 the UK government announced that nuclear weapons activities had stopped. Gas-cooled designs were also seen as being inherently safer than contemporary WCRs. A distinctive feature of the current generation of UK reactors, known as advanced gas-cooled reactors (AGRs), is they operate at higher temperatures than WCR types.
In a typical AGR, the gas exiting the reactor core can reach 640°C and is passed through an array of internal boilers to raise it to high-pressure, high-temperature steam. As a result, AGRs achieve better efficiency than other reactor types in converting the heat generated by the nuclear reaction into electricity, but they are also more complex and require heat-resistant materials. The UK is the only country that operates AGRs so any problems with them must be solved independently.
EDF Energy owns and operates the UK’s total fleet of AGRs, comprising 14 reactors across seven twin-unit sites, which were connected to the grid during the 1970s and 1980s. Almost all of them have had their operating lifetimes extended. Extending a reactor’s lifetime involves demonstrating to the Office for Nuclear Regulation that the reactor is guaranteed to safely operate over a given period.
Several key assemblies in each reactor are non-replaceable and slowly degrade over time. These include the graphite reactor core and the boilers, which use hot CO₂ exiting the core to raise steam. In an AGR, the boilers are located inside the reactor’s pre-stressed concrete pressure vessel and are not easily accessible.
To complicate matters further, major design variations exist across the AGR fleet. There are effectively four AGR generations, ranging from the Dungeness B plant built 1965-1983, to the later Heysham 2 and Torness plants. The design of each power station was refined in response to construction and operating experience of the former.
Among the areas of interest are the structural condition of each AGR’s graphite core. AGR reactor cores are constructed of non-replaceable blocks of moulded graphite manufactured from Gilsonite, a type of solid bitumen mined in Utah, USA. This form of graphite was selected and developed specifically for AGR use, to achieve good mechanical properties and minimise warping during fast neutron irradiation. There are around 675 tonnes of graphite in each AGR core and at least this much again in the surrounding neutron reflectors and shields. The graphite is exposed to neutron irradiation and high-temperature CO2, which causes it to slowly degrade over time due to radiolytic oxidation, causing mass loss, warping and cracking.
Graphite moderator blocks have vertical channels for fuel assemblies and control rods. These are the reactor’s main shut down mechanism so any movement of the core that prevents rods being inserted would be unacceptable. That said, defence-in-depth is a major principle of nuclear safety – AGRs can be shut down by injecting nitrogen gas or boron-containing glass beads into the core to absorb neutrons and slow the reaction.
As well as guaranteeing the integrity of the core during normal operation, each safety case for continued operation of an AGR considers the risks posed by events outside the power station. While the UK is not known for severe earthquakes, even small seismic events are key considerations for AGR core integrity. To allow thermal expansion, the core moderator blocks are not rigidly bonded to each other but held together by graphite keys fitted into slots cut into each block. The whole assembly is supported within a core restraint structure but the individual moderator blocks can move slightly.
Cracks in the graphite moderator blocks at the Hunterston B power station were reported in 2018. Moderator block cracking were predicted and monitored, and AGRs could be operated safely with some moderator cracks. But the Hunterston B cracking was more widespread than in the rest of the AGR fleet. Reactors 3 and 4 were shut down after planned inspections were carried out in 2018. At the time of publishing, EDF Energy had restarted Reactor 4 at Hunterston B, while approvals for Reactor 3 were still pending (see page 6).
EDF Energy has worked with the University of Bristol to check whether extreme seismic events would not cause the core to move in a way that could prevent the reactor from shutting down safely, even if the core contained cracked moderator blocks. This research has involved a combination of computer modelling and tests using physical models. For example, at Bristol’s unique Earthquake and Large Structures (EQUALS) facility, a quarter-scale model of the core with thousands of embedded sensors has been tested on a shaker table, which can simulate severe earthquake conditions.
Using results from the shaker table experiments along with computer modelling, EDF has been able to show that even in a one in 10,000-year seismic event, the distortion of the core channels in Hunterston B Reactor 4 would still allow its control rods to move freely. This was a key part of the safety case, which justified the restart of the reactor earlier this year.
Checking the condition of boilers and pipework inside reactor pressure vessels is challenging. Several different grades of steel, including 9Cr-1Mo ferritic steels and Type 300-series austenitic stainless steels, are used in these structures. Many of the metal parts inside the reactor operate at a temperature range that allows the metal to slowly deform over time when stressed, or to creep. Also, repeated cycles of reactor startup and shutdown cause a form of combined creep-fatigue loading. Many of the parts involved are metallurgically complex, for example welds between different metals.
Assessing the integrity of structures under these conditions is difficult. In response, EDF Energy and its partners have developed the R5 assessment procedure to predict the lifetime of high-temperature components based on design, material and plant data.
Methods used to guarantee parts must always err on the side of caution. For instance, a high-temperature part’s usability is normally determined by what is known about its creep behaviour than by the creep behaviour itself. Recently, high-temperature mechanical tests on boiler materials showed that their actual rates of creep under reactor conditions are lower than predicted because previous estimates used simple, cautious calculation techniques. Armed with new information and assessment methods, engineers can now make better estimates of the safe life of AGR components.
Although an AGR boiler is shielded from direct neutron irradiation, it is still mechanically stressed and exposed to high-temperature CO2. Prolonged high-temperature exposure of stainless steel boiler components in AGRs has been observed to cause the formation of secondary phases within the metal’s microstructure. These can reduce the material’s creep resistance, so the rate of secondary phase formation must be considered.
Steel parts carburise. Over time, carbon diffuses into the material from the surrounding hot CO2, which causes over-hardening of a surface layer and reduces the material’s creep-fatigue life. AGR-style carburisation is difficult to investigate in-lab because it occurs over a long time and is affected by the minor impurities, for example carbon monoxide, methane and water, present in the coolant gas. Some 9Cr-1Mo steel components are vulnerable to breakaway oxidation where the material’s surface oxide layer, which normally protects it from further corrosion, does not form properly.
Research at several UK universities, including Bristol, is being used to understand the ageing, carburisation and oxidation phenomena that occur in AGRs and determine their effects on the material’s creep-resistance. At Bristol’s Interface Analysis Centre, for example, researchers use advanced microscopy to investigate metal samples from AGRs as well as samples artificially exposed to a range of similar conditions. This enables them to study changes in the material’s microstructure as ageing progresses and hence predict its future characteristics.
Research has shown small variations in the chemical composition of austenitic stainless steel used in pipe bifurcations can produce large differences in its creep lifetime. Similarly, researchers from Bristol, Oxford and Loughborough have collaborated to develop models, which are able to accurately predict the loss of thickness in 9Cr-1Mo steel under AGR conditions.
Understanding material behaviour
In addition to operational and routine maintenance costs, EDF Energy is investing around £600mln per year on upgrades to AGRs to enable life extension. Decommissioning assets such as experimental labs and reactors across UK’s nuclear sites costs around £3bln annually, while the new Hinkley Point C power station is projected to cost roughly £20bln financed over 35 years.
The UK’s large research ecosystem in nuclear-relevant areas of engineering and materials science, as well as sectors such as chemistry, physics and geographical sciences, is a major factor in making AGR life extension cost-effective. The remaining AGRs are scheduled to operate until 2030 and continue to provide a significant portion of the UK’s base-load electricity. This helps offset the cost of building new electricity infrastructure and reduces the use of fossil fuels. However, the technical limits to the lifespan of AGRs are unavoidable and distinct from other types of nuclear power reactor.
Moderator graphite and boiler materials are not the only challenges in AGR life extension, but understanding the long-term behaviour of materials and how this affects their structural integrity is key. Like AGRs, future nuclear power reactor designs are expected to operate at high temperatures and although commercial fusion power is a remote prospect, fusion reactors have similar materials requirements. It is possible that technology being developed today for AGR life extension may be inside new high-temperature reactors in the future.
Harry Coules is University of Bristol Structural Integrity Lecturer.