The decommissioning of one of Sellafield’s most hazardous buildings has taken a big step forward with the arrival of one of three machines that will scoop out its radioactive contents.
The Silo Emptying Plant (SEP), a “kind of giant fairground grabber machine on wheels,” will take 30 years to remove radioactive waste from the Magnox Swarf Storage Silo (MSSS), an ageing storage plant which contains a quarter of Sellafield’s intermediate-level waste.
The concept of building a machine to grab the waste was agreed in 1992, and after a lengthy and complex design and manufacturing process the first parts of the 360-tonne plant are finally arriving on site. A 50-tonne ‘transfer tunnel’ was the first to arrive in November, ahead of being hoisted into place in the MSSS building. The tunnel is the main component of the first SEP, and in the coming months all its 22 modules (containing 179 plant items) will be delivered by road from Wolverhampton to the site, where the bespoke machine will be assembled by staff from Sellafield and contractor Ansaldo NES.
The SEP works by travelling along giant rails at 20m per hour, locking onto the silo hatches, and lowering its grab into the 16m-deep waste compartments through a small 1.5m2 aperture. It then brings up the waste, packs it into nuclear skips and transfers it to one of the site’s stores where it will be kept safe and secure until a decision has been made on a long-term storage solution.
As key figures from the project will attest, designing such a machine to work safely and consistently for 30 years within a challenging, space-restricted and highly radioactive environment has been no easy feat.
Chris Halliwell, head of programme delivery at Sellafield, explains that over the years the silo has been added to as more waste storage space was needed, creating unique decommissioning challenges.
MSSS is made up of 22 concrete compartments, each 16m deep and with 7m x 7m cross-sectional area, into each of which Halliwell says could be fitted five or six double-decker buses. The first six of the silo’s compartments were built in the 1960s, with the final extra capacity added in the early 1980s.
Explosive issue
Halliwell says that the silo has been filled with “all sorts of waste,” predominantly from the original Magnox swarf reprocessing programme. This waste mainly consists of the stripped magnesium alloy cladding (swarf) from the original 1m long uranium metal spent fuel bars that were sent to Sellafield for reprocessing. The magnesium alloy was stripped off the bars which then went to be dissolved and reprocessed. The remaining magnesium swarf was transferred to MSSS and emptied into compartments using heavily shielded tipping machines and maintained under cover of water.
Halliwell says: “The principal issue we have with the Magnox material is that it is pyrophoric – it ignites spontaneously in air – which is why we keep it under water.” However, when under water the magnesium reacts to generate hydrogen. “Either way you’ve got a problem. It’s fire in the air and hydrogen under water.”
To prevent the silo from generating explosive levels of hydrogen in the space above the waste, staff must ventilate each compartment.
In addition to storing 600 tonnes of swarf, MSSS has been used for various other waste materials from around the Sellafield site and other nuclear programmes in the UK. This includes “quite a bit of material from the nuclear submarine programmes that would appear overnight in the 1970s and 1980s,” says Halliwell.
He adds: “We don’t exactly have a full inventory of what’s in there – we’ve got a pretty good idea but we do have to plan for surprises.”
Other safety concerns revolve around activity to reduce the volume of waste in silos in the 1970s that resulted in some cracking to some of the original buildings and leaks to the ground. Halliwell says this issue has healed itself without intervention but there is potential for the cracks to reopen and for leakage to the ground during the SEP waste retrieval programme. “It’s a low probability but one we have to be aware of,” he adds.
A great deal of care has been made to protect people working in the SEP from radioactive doses from the waste, which has seen the teams increase shielding over time. However the predominant issues arose over increased understanding of the reactive nature of the materials in the silo.
George Andrews, chief operating officer at Ansaldo NES, manufacturer of the SEP machines, explains that there were two significant safety case developments that had a big impact on the design and construction of the machines. First, learning of the pyrophoric nature of the waste materials and the fears of hydrogen release. Second, the discovery of waste, such as large contaminated engineering debris they previously didn’t know the silos contained, leading to the addition of new handling tools. “These have meant the machines have been built, stripped down, redesigned, rebuilt and repeated all over again as these waste characteristics emerged,” he says.
Simple prototype
The original SEP specifications were determined 15 years ago and based on a Mark 1 prototype retrievals machine that successfully recovered waste from the third extension to the Sellafield facility in the 1990s.
Paul Quinn, engineering manager at Sellafield, says that the initial SEP design was a really simple system: a shielded steel cave with a hoist module outside next to a hose module that provides hydraulic power. The whole cave was based around maximising throughput and the fact that space was at a premium. The skips and transport flask, once shielded, could not go over 50 tonnes.
However, when the first safety issue came to light concerning the pyrophoric nature of the materials and the potential release of hydrogen during removal, the design needed to be altered.
Andrews says one step was to ensure that the waste could be retrieved in an oxygen-depleted environment. “With a design that was already challenging in terms of its available space, having to retrofit loads of pipework for nitrogen to be injected and create this oxygen-depleted environment was a pretty challenging task,” he says.
It was also feared that when the SEP grabbed waste it would collapse into the hole just dug and potentially create a significant and instantaneous hydrogen release and lead to the risk of explosion.
To overcome this a concept called flat-bed retrieval was developed. “This idea cropped up 10 years or so ago,” says Andrews. “It sounded great in principle. It meant that you’d deploy a ‘rake’ through the charge tube in the top of the silo using the SEP machine. It would pull waste into the centre and in essence keep the waste bed reasonably flat as you progressively reduced the level inside the silo.”
While the principle was relatively simple it required the Ansaldo team to completely strip down the SEP and design a much more complex system called a First Stage Deployment (FSD) tool. This included complex lead screws, positional tracking devices and then retrofitting it to the SEP. Andrews describes the process as a bit like putting a Swiss watch inside a battleship.
To follow on was the Second Stage Deployment (SSD) tool. This allows the waste grab to be positioned more precisely.
The mechanism to collect the waste was a basic petal grab that hangs off a wire rope and is lowered into the silo compartment, but this could only be traversed around the compartment aperture. Quinn says they knew it wouldn’t be reliable enough for the 90,000 grabs and complicated manoeuvres expected to be carried out over the lifetime of the machine.
This was confirmed after the rope failed an endurance test after only 5,000 cycles, says Quinn.
The hoist and crane system became a lot more complicated when the wire rope needed to be deployed through 90° turns. Quinn says: “What was really difficult was the cave was made and we now had to get this mechanism that would move this rope around. We went back to the drawing board and created the FSD.”
The team launched a programme of development with Reading Rope Research to design specialist wire ropes to withstand the tight bend radii in SEP and be suitable for deployment in the challenging radioactive environment.
The National Centre of Tribology also helped with the FSD to create a wider range of movement for the hoist system.
The option Sellafield came up with was two parallel lead screws moving a carriage that would push the rope backwards and forwards and left to right.
“We thought it was quite simple – a couple of lead screws run housed in some through wall plugs and we’d keep all the drives outside, which is a principle we’ve always stuck with,” says Quinn.
However, the tribology centre identified problems with the PV curve and recommended that the team find an alternative to the brass nuts they were using, which would wear incredibly quickly. Instead the team ended up using plastic nuts running on a stainless steel lead screw, which has a wear rate of 5:1. They plan for the system to have a 25-year life, with only one replacement of the plastic nuts expected to be needed. A major benefit is that the whole system runs without any need for maintenance or lubrication, reducing the risk of failure. Given the radioactive conditions, it is valuable to avoid the need for remote maintenance.
Maintenance complexity
The skips that transport the nuclear material are run up and down an 8m tunnel by drives pushing a bogie backwards and forwards. This avoids the use of motors or drives in the radioactive environment that would need much more complex maintenance if they broke down, explains Quinn.
The design uses an off-the-shelf Serapid push-pull chain that the team knew to be robust, yet it failed during the endurance test. This was because the high carbon content of the steel chain pins and poor quality bushing material rusted and seized up in the simulated environment. Quinn says: “Every aspect of what we’ve done is to start off with a really simple premise that we know works, but when you shove it into this really tight space with its constrained, non-maintainable, really high alkaline, high radiation environment it gets difficult and you need to get more technical.”
Sellafield turned to the National Centre of Tribology to advise them on new material choices, and a nickel undercoating for the chain pins was suggested. “Chrome is a nice hard surface finish and runs with the new bush that we selected, but it has tiny microscopic cracks in it that would still allow corrosion,” explains Quinn. “A nickel-based coating underneath gives you corrosion resistance and then the chromium coating on top gives you the required hardness.”
However, to achieve this was no easy task. They would first have to buy the off-the-shelf chain from Serapid, then send it to a contractor to strip the chain down, send the pins overseas to get the nickel coating and then back to the UK for its final chromium coating.
Cutting by water jets
Perhaps one of the biggest design challenges, says Quinn, was the discovery of engineering debris in the silo’s compartments, namely 15 huge swarf bins that were used to tip waste into the compartments back in the 1960s and 1970s.
The bins were found after the SEP was built and it was realised that they wouldn’t fit into the shielded waste skips. They would have to be cut to allow them to fit inside. But aggressive cutting tools that would give off sparks could not be used because of the hydrogen risks. The team opted for a high-power water jet system but it was too powerful to be held by the manipulator, which would lose control and could risk going through the cladding and causing a containment issue.
Instead the team created a special skip which pulls up the swarf bin so that a nozzle set in a turntable can spin and cut the bin with the water jet into a manageable size, explains Quinn. The bins, however, are often full of sludge and can weigh 1,800kg. During tests these items were picked up with a 50% success rate. If the bins were to hit the walls of the SEP cave that could potentially be a very expensive write-off (each SEP costs
£80 million). This led to the addition of more reinforcements to the steel cave to counter these droploads.
A device called a ‘defamator’ was also added to the SEP. Halliwell says this is akin to hydraulic shear packs that firemen use to cut people out of cars. This can twist long items so they will fit in the waste skips. All additional tools can be deployed with the electronic manipulator which is operated by workers to support remote operations and maintenance.
Andrews says such solutions are required because an action such as withdrawing a plug or a shaft to change something as simple as a limit switch becomes significantly more difficult in a radioactive environment. He adds: “You need to design a whole suite of maintenance tools to allow you to withdraw fairly simple equipment but in shielded containers.”
The SEP machines will have to be assembled and work in very tight spaces. This meant Ansaldo had to modularise the design so that no one piece was more than 50 tonnes so it could be lifted through the building’s hoist well, with just 40mm clearance on either side for the largest module. This means each part’s centre of gravity has to be absolutely perfect, says Andrews.
This has meant extensive use of 3D CADCAM and a lot of “practical trial and error” to find the centre of gravity and adjust and redesign lifting attachments so there is absolute confidence they will fit.
The parts must then be reassembled inside the silo. It is a very congested facility, explains Halliwell, with significant amounts of other project work going on around the SEP such as a new ventilation system, and is “a bit like a Chinese puzzle”. However, Ansaldo has already carried out its own build and testing of the first SEP and the firm is confident it will be able to transfer its know-how to the construction team at Sellafield.
Andrews says that the second machine is coming towards the end of its build sequence and all of the learning, experience, modifications and challenges that were identified during the building of the first machine, which is in essence a prototype, have been incorporated. The third machine will likewise be further optimised.
There will now be a nine-month reassembly process for the first machine during which more lessons will undoubtedly be learned. The first waste retrieval activity will begin towards the end of 2017 in what is Sellafield’s most technically challenging programme. Quinn, at least, is not phased by the work ahead. “It is exciting,” he says.
Did you know? The Silo Emptying Plant
The plant works by travelling along giant rails at 20m per hour, locking onto the silo hatches, and lowering its grab into the 16m-deep waste compartments through a small 1.5m2 aperture. It then brings up the waste, packs it into nuclear skips and transfers it to one of the site’s stores where it will be kept safe and secure until a decision has been made on a long-term storage solution.
The Silo Emptying Plant, a “kind of giant fairground grabber machine on wheels,” will take 30 years to remove radioactive waste from the Magnox Swarf Storage Silo