The decommissioning of the fateful Windscale reactor at Sellafield may not be as hazardous a task as experts have always believed. George Coupe reports.
‘There’s a joke that’s going around Sellafield,’ says Paul Worthington. ‘If anything happens outside all we have to do is come in here.’ ‘Here’ is a decontaminated ante-chamber within the 2m-thick concrete bioshield of Windscale’s number one reactor at Sellafield. It was this reactor that caught fire in 1957, and until Chernobyl the incident ranked as Europe’s worst nuclear accident.
The ‘joke’ is that, thanks to a successful clean-up operation around the fringes of Pile One, it would now be safer to be inside this section of the bioshield than outside should another accident occur at the adjoining BNFL site – a little laboured perhaps, but it makes the point.
We are not, of course, standing within the reactor tank itself. This has remained sealed since the accident and still contains about 15 tonnes of uranium fuel. Nobody has touched it for almost 50 years because of a fear that it could either catch fire again or go critical and explode.
Paul Worthington is head of safety and environment for the United Kingdom Atomic Energy Authority (UKAEA), the body charged with decommissioning the reactor. It would be hard to see how the problems he faces could be much worse. But, standing at the charge face of the reactor, just yards from the worst nuclear mess in western Europe, he is surprisingly optimistic. The reason is that he and his team have just made an important discovery, which could make decommissioning much easier than was previously thought.Not many people remember what happened at Windscale in October 1957. The remote site on the Cumbrian coast is where the UK secretly launched its nuclear weapons programme in the late 1940s.
The US had refused to let the UK share its atomic weapons technology, so the decision was taken, without reference to parliament, to build two massive air-cooled reactors at Windscale to produce plutonium for the British atomic bomb. Building work started in 1947 and, after only three years, Pile One started operation. The process involved irradiating large amounts of uranium fuel to create weapons grade plutonium.
Each core consisted of 2,000 tonnes of graphite blocks arranged in an octagonal stack, 16m in diameter and 8m long, surrounded by a biological concrete shield. The graphite blocks were shaped and fitted together to create 3,444 horizontal fuel channels, and 909 isotope channels through the core. The fuel was loaded from the charge face – 180 tonnes of uranium was needed to achieve critical mass. In the intense heat and neutron flux a small amount of the uranium was converted to plutonium. The irradiated fuel was then pushed through the reactor and fell out at the discharge face into a water duct, where it was carried away to be processed.
Graphite moderates the nuclear reaction but its major drawback is its tendency to store latent heat energy which it then releases suddenly and unexpectedly. Known as Wigner energy, the only way to control this is to heat the graphite core to achieve a controlled release. On 8 October 1957, during the ninth heating or annealing of Pile One, faulty temperature-measuring instrumentation gave control staff a false impression of the heat level inside the reactor. Thinking that it was not hot enough, they gave it an extra boost. As the control rods were withdrawn temperatures soared, and some of the fuel and isotope material at the centre of the reactor caught fire.
The massive air-cooling system, capable of pumping a tonne of air every second into the reactor, fed the flames and sent radioactive iodine, plutonium, caesium and polonium contamination up the giant chimney where the filters were unable to hold it back. A cloud of fall-out drifted across south Cumbria towards the cities of the north of England. Desperate to stop the fire and prevent an explosion, staff used scaffold poles to knock unburned fuel out of the reactor. Eventually, after two days and at great risk of disaster, the fire brigade used water to douse the flames.
Afterwards Pile One was sealed up and has remained so ever since. Pile Two was immediately shut down and defuelled. However, a large amount of fuel, about 6,700 fuel rods (11-15 tonnes of uranium), still remain in Pile One, with the potential to catch fire or go critical once again.
Pile One has stood as a giant black mark on the record of the nuclear industry, and been viewed over the years as too risky to clean up. But after extensive research and analysis, Worthington is confident UKAEA can have all the fuel and isotopes out of the reactor within the next 10-12 years. This could bring huge benefits to the nuclear industry. The sooner Windscale is dealt with the easier it will be for the country at large to contemplate a future with nuclear power. In the period immediately after the accident, between 1958 and 1961, the area around Pile One was decontaminated. Control and shut-off rods were inserted and locked in, and the operating gear removed. The pile cap was sealed to prevent further damage to the core and the blowers and air filters were removed, and the air ducts sealed.
In the early 1980s UKAEA fully sealed off the core. The water duct was drained and shut off; a seismic-resistant barrier was fitted to the giant air inlet and a new ventilation and monitoring plant to detect disturbance in the core was installed. North Sea exploration expertise and equipment were required to clear the water duct of old fuel elements, contaminated silt and graphite ash, and military robots were used to retrieve stray fuel elements from areas such as the exhaust air ducts. The work was finished in 1999.
Phase two clean-up began in 1999, with the objective of dismantling Pile One as soon as possible. Pile Two would be put into ‘surveillance and maintenance’. A contractor was appointed and given a choice of four methods to explore. The first option was to fill the bioshield with water, and then dismantle the core, using standard storage pond techniques. The second was to freeze the core, and then take it apart in air. ‘This would hold the core together, provide cooling but also isolate the core from air which is very important,’ said Worthington. The third option was to dismantle it in air, but with argon on standby to fight any fire that was expected to break out. The fear of another fire stems from the assumption that the core now contains a considerable amount of uranium hydride. This would have been created when water was poured on to the reactor in 1957. When water is poured on to hot uranium, uranium hydride is produced, which is pyrophoric.
‘The postulated scenario was that there might have been a tiny hole in the aluminium fuel canisters,’ said Worthington. ‘Water got in when the fire was put out creating the hydride, and for whatever reason the hole was sealed, trapping the hydride inside, isolated from air. ‘If you do something to disturb that fuel element and open up that hole again the air gets in and the hydride reacts spontaneously and generates heat and there is possibly another core fire.’
It was thought that argon could be used to isolate the core from air and prevent the hydride oxidising. The fourth option was to fill the entire bioshield with argon. This was the method chosen by the contractors, a consortium of BNFL, Rolls-Royce and Nukem.
After establishing an argon-inert atmosphere within the reactor, manipulators mounted on the pile cap would reach down into the core and pick and place the pieces of graphite into baskets which would be taken off to a treatment plant and store.
‘But when we started to get into the detailed design [in 1999] we started to identify problems,’ said Worthington. ‘The manipulators could not reach everywhere they had to. It was not possible to put more holes for additional manipulators in the pile cap because that weakens the structure. The argon environment was an industrial safety issue [argon is an asphyxiant] and the volume required was equivalent to half of Europe’s production.’
In 2001 UKAEA decided that a new approach was needed. Instead of a turnkey contract it opted for phased design, development and implementation of decommissioning, managed by UKAEA. The new approach was influenced by a process developed by the US Department of Energy. ‘What we had done before was to start with a scheme (for decommissioning) and detail it up, but what we needed to do was go back a stage and identify the hazards first, and the event sequences that might lead to a disaster.’
Pile One can, it is thought, harbour a wide range of serious hazards. Apart from the uranium hydride reaction, there is the potential for the remaining 15 tonnes of fuel to go critical. The fire may also have damaged the graphite so badly that the structure is in danger of collapse, which could lead to a fire or criticality. There is also the possibility of a graphite dust explosion. It is assumed that the fire produced a certain amount of dust within the core, which could explode and breach the bioshield.
First step in the new design process is to assess all these hazards. The team will work on five plans with a view to picking a favourite by 2005. The aim is for a strategy that will deal with as many of these event sequences as possible in one go. For example, grouting the core, filling it with concrete, would prevent the uranium hydride reaction and also prevent a core collapse. Once grouted the core might be cut up with diamond wires. Other options include supporting the core from collapse with steel bars inserted into the empty fuel channels. To prevent criticality, an absorber or poison could be added to the core, such as boron dust, to prevent an explosion if the fuel came together.
Worthington said it might not be necessary to dismantle the core completely. Grouting would go a long way to eliminating most of the hazards that it currently poses. An alternative might be just to remove the fuel and isotopes and seal up what is left. The entire core could eventually be removed at a later date.
At this early stage the different strategies are ranked in terms of safety, effectiveness and technical feasibility, but not cost. ‘If you put that in too early you get driven to a cheaper solution, which is not the way the process is meant to work. Cost and programme come later,’ said Worthington.
Whatever is decided, no operations are allowed in or around the core without approval from the regulator. Activities at the plant are governed by the ‘operational safety case’, which is based on long-standing assumptions about the state of the pile and how it could catch fire again. As part of the decommissioning process the OSC is being updated, and new thermal modelling techniques have now revealed that some of these conditions, which make Windscale the highest risk category of plant, do not actually exist, according to Worthington.
The discovery, he said, was produced by an attempt to model how a fire might propagate through the core. The experiment was fed with technical expertise from the US and the latest research into uranium hydride reaction rates by BNFL. Instead of resulting in a major conflagration, the model said the opposite would happen.
‘The more we got into the study the more we realised that there was sufficient technical evidence that a core fire could not start now. You might get uranium hydride that reacts in air and heats up. But the modelling shows that the temperature rise from ambient temperature is too low to start uranium burning. Even if the hydride is attached to the uranium metal itself it can’t warm up the uranium enough to start a fuel fire, it can’t get the graphite hot enough to make it oxidise, and it can’t get from the fuel channel to the neighbouring isotope channel.’
At a stroke the task of decommissioning Windscale Pile One had been transformed. ‘We are saying the uranium hydride to core fire hazard which has been hanging round this project since 1957 really isn’t there,’ said Worthington.
The research is going through peer review at present and UKAEA hopes to present it to the Nuclear Installations Inspectorate later in the year. If the hypothesis is approved it means the NII will have to agree there is a much reduced risk of contamination within the core escaping to the outside world.
‘The importance for us is that instead of being a category one plant , which is the highest risk category, we are now potentially looking at being a category three plant. And the decommissioning strategy becomes much simpler. We will not have to isolate the core from air, which means we won’t need to use argon to cope with the metal hydride reactions. It is now thought that this could allow us to get the fuel and isotopes out of the reactor within the next 10-12 years.’ The next step is to make some more detailed surveys of the core.
In December, pictures taken of the discharge face revealed that many of the channels in the fire-affected zone are now blocked with debris, including molten aluminium fuel canisters, burnt uranium fuel and graphite ash. The images help to determine roughly what temperature the fire reached, which in turn will help the team to estimate the damage to the graphite structure and the likelihood of a core collapse.
So far the investigation showed that the front and back face of the core are sound. The graphite blocks do not appear to have been disturbed. A camera put into two ‘foil’ holes that run from the top to the bottom of the core also revealed the blocks were still perfectly aligned – though these were at the extreme ends of the reactor.
Worthington hopes that discounting the possibility of a core fire will make it easier to get approval to put a probe down a foil hole directly into the fire-affected zone to make a full assessment of the damage. However, the question of ‘criticality’ has already been similarly simplified using computer models. The core is acceptably sub-critical at present and in numerous collapse scenarios that margin remains. Worthington said the most likely route to criticality will be an error in the fuel and isotope removal sequence.
If UKAEA can remove the fuel and isotopes as soon as it hopes a question still remains over what to do with the thousands of tonnes of graphite from the core. Graphite is a common core material and nuclear waste storage is a precious national resource in the UK, and will remain so until a deep repository is built. As more graphite core reactors are decommissioned over the next 10-15 years, a national strategy for graphite disposal will have to be devised.
‘We’ve either got to store it or come up with some means of getting rid of it. It’s not just us. We’ve got 4,000 tonnes, Calder Hall has graphite reactors, there are lots of other graphite reactors around the country so it’s a national issue,’ said Worthington.
The UKAEA is experimenting with a method of treating graphite with low levels of contamination that could allow it to be disposed of in landfill. Tests on graphite from GLEEP, the experimental reactor at Harwell, which has almost been completely dismantled, have shown that putting the blocks in an incinerator dramatically reduces their radioactivity while they do not catch fire. UKAEA is now carrying out similar incineration tests on crushed graphite blocks in the hope that this will reduce levels of radioactivity further.
At Windscale Worthington said the fuel and isotopes from Pile One could be handed to BNFL for storage, while in return UKAEA looks after the graphite from BNFL’s Calder Hall power station. Such economies of scale built into the plan for Windscale could, Worthington hopes, find favour with the newly established Nuclear Decommissioning Authority (NDA), responsible for prioritising funding for the clean-up of the UK’s nuclear legacy.
Once the fuel and isotopes have been removed, said Worthington, the graphite core could be left inside the bioshield which would be sealed up until approval is given for a national strategy for graphite disposal. Then only one treatment plant would be built to take the graphite waste, 4,000 tonnes from Pile One and Pile Two, and the rest from Calder Hall.
Final decommissioning of Pile One is set for 2037 under UKAEA’s latest accelerated decommissioning plans, but there are many factors that could delay that, including the time it takes to get approval from the various regulatory bodies.
The crucial first step towards removing Pile One will be approval from the NDA, which is soon to accept bids for funding. Worthington believes Windscale has the X-factor, which will come into play when the NDA decides on its priorities. The plant is a highly visible reminder of all that is generally considered wrong with nuclear energy. UKAEA hopes the new assessment of Windscale will offer the NDA an opportunity for an early success.
It will also help to prove that nuclear expertise, knowledge and safety have come a long way since the day of the fire in 1957. ‘Even with a plant in this condition we can show we can deal with it. This then should add confidence and show that the nuclear industry can get a handle on itself,’ said Worthington.