At the UK’s largest nuclear complex, Sellafield, in north-west England, decommissioning of the oldest buildings is complicated by the fact that the original 60-year-old facilities were not designed with any regard to future decommissioning. The site is very crowded with little room to build next to the legacy buildings. Furthermore, insufficient treatment and storage facilities are available for existing wastes.

The project to decommission an intermediate level waste storage silo faced all of these issues, and more. The pile fuel cladding silo (PFCS), Figure 1, is a tall (24 m high), relatively narrow box of a building set back from one of the busy roadways of the site lined with cabling, ducts and pipe-bridges in the centre of Sellafield, where legacy buildings jostle for space. It was built in 1950 to receive solid waste arising from the decanning of fuel used in the two Windscale graphite reactors (piles) built to produce plutonium for the UK weapons programme, which went critical in 1950.

The 1950s Windscale plant was built in haste as the UK rushed to develop a nuclear deterrent. Many of the buildings were first-of-a-kind, for which no design or building regulations existed. The PFC silo was modelled on a Canadian grain silo and built to the engineering standard of the time. The original specification has been lost, but it is known that the silo was intended to hold intermediate level waste from the decanning of low burn-up fuel in the two piles. It operated from 1952 to 1965, with occasional tipping taking place for three years until 1968. It also served to store waste from the early magnox reactors as well.

Both Windscale and magnox reactors were originally designed to use natural unenriched uranium as a fuel. The uranium was encased in canisters, or cans, made of aluminium (at Windscale) or magnesium alloy (magnox). The fuel rods had fins to help radiate heat. Spent fuel rods were run through a machine to strip off the cladding and fins, and packed into high-level waste containers. The remaining waste found its way to the silo.

The silo consists of six compartments in a reinforced concrete structure (Figure 2). Flasks of waste were delivered to the facility by road transporter and were lifted up the external hoist-well to one end of the roof of the plant. A bogey transported the flask into the antechamber of a steel and brickwork transfer tunnel where its lid was removed. A second trolley trundled the flask down the transfer tunnel where it was positioned above the charge-hole for one of the six compartments. The flask was then inverted, depositing the contents down the hole. A deflector plate below the charge-hole was intended to spread the waste out as it fell into the silo and prevent excessive mounding. The original design provided each compartment with sumps with sand filter beds and a drainage system.

Care and maintenance preparation

As regulatory requirements increased, two significant upgrades were made to improve the containment and safe storage of the waste in the 1970s and 1980s. Further shielding was added to the building to reduce the external dose rate which was affecting work on neighbouring plants. The extra shielding consisted of some slabs bolted to the external walls, and some free standing shield walls around the building. Some of this was done during the operational life of the plant and some was done later.

A second modification, made during the 1980s, was to fit fire detection and fire fighting systems enabling the injection of argon gas at both top and bottom of each compartment (see also box).

More recently, the silo has been the subject of a programme of building improvements to support retrieval operations. Inspection of the external silo walls showed limited corrosion with no significant repair necessary. The structure was reinforced with carbon fibre straps to meet future dynamic loading during retrieval. The walls were coated with a sealant to enhance argon retention.

A new staircase leading to the roof was fabricated and work carried out on the roof structures. The reflector plates at the top of the compartments caused some trouble as waste had built up on them and backed up into the transfer tunnel. Clearing this enabled the charge holes to be blocked, the silo sealed and the transfer tunnel demolished by 2003. The secondary shield wall surrounding the silo has been removed to allow access to building. This involved the dismantling of 325 two-tonne concrete blocks and 65 tonnes of steelwork, which were then free released. Significant corrosion was found in the roof structure and some remediation has been necessary.

The reactive nature of some of the waste—notably uranium hydride—led to the decision to empty and then decommission the building. However, until a waste retrieval facility was built, there was no prospect of emptying the ageing silo. By 1996 BNFL (British Nuclear Fuels Limited, which owned and managed the facilities) had designed a programme to refurbish the silo and to construct waste retrieval and treatment facilities, which would then permit the silo to be decommissioned. Fifteen years later the twin constraints of finance and safety regulations mean that progress has been slow, and it is only within the last year that the pace of work under the new Sellafield Ltd management team has stepped up.

Accelerating decommissioning

Paul Nichol, PFCS facility manager, explains: “Our current strategy is to get the material out and then store it. We can’t afford to leave the silo sitting there. It is an old structure and we can’t qualify it [from a safety point of view]. We have done our best to get it into good condition, but it is an intolerable risk and does not meet modern standards.”

A major breakthrough in the plans for retrieving the waste has come from a strategic change in the site approach to risk assessment, particularly for machinery. Graham Young, PCFS operations delivery manager, said: “Our old methodology for cranes was to assess the crane for a lift of 360-degree radius. Under the new methodology we assess the risk for where the lift is being carried out.” Because the silo is so close to one of the legacy ponds, making a safety case under the old regime (even though the crane would not be lifting near the pond) could have taken nine months of paperwork. However, the highly-qualified drivers in the crane industry, and the work area limitation systems in modern cranes’ control and instrumentation software, have enabled a change in risk analysis.

Assessment work under the new regime took a day and a half. Graham Young said, “It got the point where we were so conservative that it was presenting more of a risk because we were so slow. We have taken a step back and looked at the balance of risk in order to find ways to move forward. It is as someone has given us the key to a set of handcuffs.’

Paul Nichol explained the new site philosophy on applying a balance of risk. “We recognise that the big risk is the facilities staying where they are. We have to get the facilities down and reduce the risk. There are all sorts of reasons not to do something, but we apply a balance of risk and a balance of magnitude to the strategy for building the silo waste retrieval facility. The original plan was to build without using cranes, because of risk of a crane falling on the silo. But we can balance the risk of the crane: properly installed, properly managed cranes do not fall down.”


The waste retrieval facility will consist of a new reinforced concrete support structure with modular steel retrieval and packaging cells that are structurally independent from the silo but adjoin it. A number of modules will be bolted on: retrieval modules; waste loading module and waste container modules. The steel structure modules will be able to empty each compartment in turn.

The silo has 12 compartments, six per long side. The five internal walls separating the compartments would block a retrieval structure covering the entire side of the silo; instead, a modular structure that can be moved along the building will be used (see Figure 3). The latter is cheaper and quicker to build; it will also be cheaper to decommission.

Six 8-metre high doors will be made, one for each compartment. They will be set onto door posts on the new structure. The doors will be bolted to the silo structure and sealed against the face of the silo. The retrieval module will dock against this, maintaining the integrity of the silo’s argon-inerted environment while waste is retrieved.

Holes will be cut into the north side of the silo so that each compartment can be approached in turn. Development work is underway on the hole-cutting programme.

Once an opening is made, a telescopic boom holding a clamshell bucket will be deployed through the new holes in the side of the silo (Figure 4). The boom will extend and the bucket will be winched down to pick up a load of material. Then the bucket will be winched up, the boom will retract, spin through 180 degrees and lower the bucket again into the waste characterisation module in the hoist-well of the new structure. The waste will be spread onto a table using a robotic arm. Radiological fingerprints, photographs and video footage will be taken to aid characterisation and treatment. When characterisation is completed, the robotic arm will push the waste into a waiting three-metre cube storage box, which will be moved to a lidding station, swabbed and exported. The boxes will be protected within a flask and exported to buffer storage prior to treatment and final storage.

As UK geological disposal is still under discussion, and no prospective sites have been identified, there is currently a lack of certainty about the design of final disposal containers in the UK.

Graham Young said, “We have to get the material out and into safe storage. We are working closely with repository staff to make sure that we are prepared.” A generic disposal container is under development and Young says, “We must ensure that nothing we do compromise that.” Three thousand three metre-cubed boxes, at a price of £50,000 apiece, are on order for the PFCS intermediate-level waste. Young describes their design as ‘an educated guess.’

The design, installation and commissioning stage of the waste retrieval facility is due to be completed by mid-2017 when active commissioning and retrieval is scheduled to commence.

Making room for the retrievals facility has been a major logistical exercise. The PFCS office accommodation has had to be shifted to the south side of the silo. On a conventional site, hiring a set of cabins would have been the answer, but amidst the legacy facilities of Sellafield it was necessary to build a heavily-encased concrete building with lead-shielded roof to protect workers from the dose rate and shine emanating from the silo in the only space available to re-site the offices. Using a crane to move the existing offices enabled the job to be completed in less than two days; without it the job would have taken six months.

As of time of writing in late April, the concrete foundation slab for the retrieval module had been finished; construction of the concrete superstructure shell started in April 2011. Substantial work is required to build the 500 ton retrieval modules and get them into place; they are too big to transport by road and are too heavy for a conventional crane to lift. A semi-goliath crane and strand jack system will be used to install them.

A Babcock and Bechtel joint venture won a £120-£150 million, six-year contract to build the structure in three phases: mobilisation, detailed design and construction.

Related Articles
Sellafield silo retrieval project enters second phase

Inerting the silo

In the 1990s a decision was taken to inert the silo in order to guard against the risk of fire. A major concern with the silo inventory was the pyrophoric uranium hydride, formed by the corrosion of uranium metal in damp and oxygen-deficient atmospheres. Development work to determine the formation rates of uranium hydride in argon-air mixtures showed that hydride formation was inhibited by trace quantities of oxygen of the order of 8 ppm. Thus, provided oxygen concentration levels were above 8 ppm and below 2 percent, the risks posed by uranium hydride could be safely managed. Argon was chosen in preference to nitrogen as it would extinguish a magnox fire better than nitrogen.
The design, installation and commissioning of two redundant argon inerting system was in place by 2001. Liquid argon is stored then converted to gas and pumped into the silo. The argon level is monitored and maintained to keep O2 levels at about 0.1 per cent. If oxygen levels are below 2 percent then a fire cannot start. At 4 percent it can spread.
Initially the system was run at lower pressure than the outside atmosphere, so that in the event of a leak air would be drawn in, rather than potentially-radioactive gas escaping from the silo. Nichol explains how the inerting system has been refined over the years. “Previously we extracted argon gas using a forced ventilation system to create low pressure. But this drew air in and increased oxygen levels. The higher the oxygen level, the more argon we pumped in. We spent years adjusting valves, tweaking this and tweaking that to keep oxygen concentrations down and maintain low silo pressure.”
Under this pressurised system, loss of argon would quickly lead to fire as oxygen concentrations could rapidly rise above 2 percent. The team eventually convinced nuclear and environmental regulators to review the forced venting, after estimating the building’s contamination levels by measuring filter contamination. Nichol says: “It is now running at positive pressure, but low pressures. Very low levels of contamination are being carried out of the building. It is the only building on the site running at positive pressure.”
The silo, now virtually gas-tight, will hold its argon for five days before the oxygen level rises to 2 percent.