It is more than half a century since uranium enrichment by centrifuge began at the UK’s Capenhurst site, replacing space- and energy-hungry gaseous diffusion technology. The thousands of centrifuges at Capenhurst – and Urenco’s other enrichment sites, Almelo in the Netherlands, Gronau in Germany and Eunice, New Mexico in the USA – have been quietly and consistently enriching uranium over the decades. But that consistency does not mean there has been no development of the technology, or discussion of the latter half of its life.
Advancing centrifuge performance
Mike Peers is Chief Nuclear Officer at Urenco Capenhurst. He says that over the years “Most of the advancements are down to the centrifuge designs improving over time. As the design of centrifuges has matured, we get a lot more productivity per centrifuge than we did from the centrifuges we started to use more than 50 years ago”.
Mostly this is down to the design of the centrifuges, which are made by Enrichment Technology Company (ETC) – a joint venture between Urenco and Orano, a French conglomerate – and bought by Urenco.
Peers says, “Performance is influenced by the size of the centrifuge: length, radius, and how fast it spins.“ He adds, “We get significantly more performance from a centrifuge than we did 50 years ago simply because they have got much bigger and faster. As a consequence they have also got more efficient and you get more enrichment for a given electricity consumption. The progression has been both revolution [change of design] and evolution, as engineers have been able to iterate the design over time to squeeze the last bit of performance out of it.”
Behind the change has been significant changes in materials technology, stronger materials and quality control in the manufacturing process, because “anything that is spinning is under stress”. Peers can’t talk about the detail of that, but he says the outcome has been “probably an order of magnitude improvement over time”.
The centrifuges’ lifetime is also much longer than expected. When the first ones were installed, Peers says, “we were rolling them out with the expectation of 10 years life. But the oldest machines at the Capenhurst site started operations in 1982 and we shut them down in 2016.”
A reassessment dated 1 January 2024 added five years to the expected lifetime of some centrifuges. Peers explains, “Our centrifuges are incredibly reliable. Whatever end date we expect to achieve they outperform it, so we do have to reassess [regularly] what ‘end of life’ might mean.” Peers says that like any mechanical system, centrifuges typically have a ‘bathtub curve’, with failures more likely at the start of life and rising towards the end, and very few over most of their lifetime. Peers says, “we see almost no failures, or very low failures – so low as to suggest we are not at the end of life for our centrifuges – certainly not for the newer models. It gives a bit of confidence to increase our expected lifetimes to the machines”.
Looking back to that first generation of machines, he says: “We see a similar level of excellent performance in our newer machines and that’s what gives us the confidence to say we can extend their lifetimes by conservative amounts, by five years.”
When it comes to maintaining that performance “there is no substitute for experience,” says Peers. The working assumption is that the machines run at full capacity and there will be no outages. “After 50 years we have experience of how to operate them in a smooth and reliable manner so we get the best out of them.”
End-of-life considerations
Even the most long-lived centrifuges reach the end of their lives eventually. With thousands in operation, there is a balance to be made to optimise efficiency. “The individual machine will fail, and we can leave that machine and the plant carries on, until there are lots of machines in the same scenario,” says Peers, adding; “eventually you have enough that replacing them will be beneficial. That’s what starts to make the case for doing some refurbishment – swapping out failed machines and replacing them with new ones to restore us to full capacity”.
Of course, he says, “you still have to look after the other components and the auxiliary systems that support the centrifuges. Uranium take-off systems all need a level of maintenance and from time-to-time refurbishment, especially in the older plants, so we do targeted refurbs as well.” He adds that after the first batch of centrifuges were finally retired: “That plant from 1982 is still running: we shut down the oldest centrifuges in 2016 but we completed a significant outage in 2024 to make some upgrades to the auxiliary systems and some safety upgrades and it is now back online”.
As in all nuclear installations, the end-of-life phase for equipment and buildings implies a very different regime from their operating life.
In its annual report Urenco said; “work on centrifuge decommissioning will be a critical enabler to the success of the capacity programme and will involve dismantling, declassifying and decontaminating all centrifuge types”. That process is directed by James MacLeod, Urenco Nuclear Stewardship’s Managing Director.
As Peers has described, over time there will be steady trickle of centrifuges that have to be stripped out and replaced, with occasional larger volumes as a result of a refurbishment. MacLeod explains that all waste – not just radiological waste – has to be handled through a set process. Waste prevention and minimisation are the first considerations and MacLeod says “creating as little waste as possible,” is key. This fits with UK and EU guidance on ‘waste hierarchy’ used by many industries, which gives top priority to waste prevention, followed by re-use, recycling and recovery, before disposing of residual waste.
In accordance with this process, the company is looking at options to re-use centrifuges parts where appropriate. “Our R&D department is in discussions with ETC, our centrifuge technology joint venture, on agreeing decontamination techniques and limits to allow the outer casings of the centrifuges (called the recipients) to be returned to ETC in such a state that they would be able to refurbish them and we would be able to reinstall and reuse them,” says MacLeod, adding “It is difficult, because you need dismantling lines capable to carefully remove the internals and decontaminate to a level that is sufficient for re-use.”
That development work is ongoing. To prepare for the process, all the sites will need additional storage facilities in the first instance to store the centrifuges followed by dismantling lines and enhanced waste treatment facilities to manage the resulting waste. “They have to have a centrifuge dismantling line and that will be a number of cells that will take the centrifuge apart and decontaminate it”, explains MacLeod. In that process, he says, “We work to separate and decontaminate component parts of the centrifuges to allow re-use or to decontaminate so they are released from nuclear regulation, and thereby decommissioned.”
The R&D team is designing the dismantling line with the key objective of re-use, MacLeod says.
Separately there are other challenges. For example, MacLeod explains, “We have a unique issue in Urenco in that we have a lot of aluminium and the treatment of aluminium globally isn’t optimised – most nuclear sites that need decontamination have steel. “We are designing a melting capability. We have gone through the concept design already, to put in place the ability to melt both steel and aluminium”.
MacLeod says that Urenco have historic experience of running a melter on the Capenhurst site, although this was decommissioned in the early 2000s.
Urenco wants to demonstrate that it takes waste seriously. But it also has to meet licence conditions that require it not to accumulate any more waste than is absolutely necessary. MacLeod says, “we have waste routes for all of our waste streams to ensure we minimise that waste. We are going through re-use for as many recipients as possible.”
He also says the Capenhurst team gained a lot of experience doing this when it decommissioned the old gaseous diffusion plant, during which “we managed to process a significant amount of metal in the last few years that has gone into conventional recycling markets.”
Closing in on decisions
The R&D department currently has a single robot cell which will now be used to start doing trials on the dismantling process, using old centrifuges.
These are “ongoing conversations”, MacLeod says, that do not depend just on whether the individual centrifuge types can be refurbished. There are geographical issues that determine whether it is currently possible to refurbish and recycle, such as security considerations, says MacLeod. He notes that “Urenco leadership is prioritising reuse of recipients over other options even if it is not as economically advantageous, just so we meet all our environmental responsibilities”.
In the longer term, Capenhurst’s planned metallic treatment capability, including the option to melt, could be made available to external customers. MacLeod intends that it will become a hub in the UK for metal treatment.
