Wylfa nuclear power plant, in North Wales, comprises two Magnox CO2 gas-cooled reactors, A1 and A2, which reached commercial operation in 1971 and 1972 respectively. Wylfa was the last of the UK’s 26-reactor Magnox fleet to be built and had the greatest output. Each reactor produced 570MWe, a big step up from the other Magnox reactors, the biggest of which (two units at Hinkley Point A), were rated at 235MWe.

At the end of December 2015 Wylfa ceased final operations at its last unit, marking the end of an era in UK generation. During its life, Wylfa produced 232TWh of electricity. The site is now in a defueling phase, which will be followed by working towards a care and maintenance state.

For many years the UK’s strategy for reactor dismantling has been so- called "care and maintenance", originally referred to as "Safestore". The Nuclear Decommissioning Authority (NDA), which owns the Magnox reactors and their sites, referred to this as its "baseline strategy". It assumes that after the reactor is defueled and outer structures with little or no contamination are dismantled, the reactor itself is "walled up" and left for around 85 years following shutdown before it is dismantled.

But in a recent review the NDA said it was time to question this long-held assumption. In a draft strategy published in February it said, ‘Whilst we will celebrate as the first few sites are made safe and secure for a long period of quiescence, it is hard to ignore the question of what comes next. Increasingly we find ourselves questioning whether the baseline strategy is appropriate as a blanket strategy for all reactors in the Magnox fleet’.

The NDA said it had carried out research, with site operator Magnox Ltd, on the implications of alternative decommissioning strategies. The results ‘support a review’.

The research found two major problems with the deferred decommissioning strategy. First, it had assumed that deferring dismantling would mean that radioactive decay over the period would allow the structures to be treated as intermediate level waste – less costly to manage and dispose of and less hazardous to the workers involved. However, it said better understanding of the implications
of radioactive decay ‘have shown us that after the long period of quiescence a large amount of the reactor waste will still not be suitable for management as LLW, despite broadly fitting into the LLW category, due to high inventories of long-lived radionuclides.’

Secondly, ‘a preliminary high level cost model suggests that as the deferral time increases, the reduction in decommissioning costs – resulting from increased worker access – is largely offset by the increased cost of preparing for and managing quiescence.’

Finally it found that even after deferring dismantling for nearly a century. ‘It is likely that remote dismantling techniques would be applied to Magnox reactor dismantling as a matter of best practice to help minimise conventional safety risks and doses to workers’, so cost savings would be strictly limited. The complexity of dismantling assets that have deteriorated over the years could in practice increase the cost of dismantling, it says.

NDA lists other risks that it weighs against the benefits of deferred dismantling, including: loss of skills, knowledge and capability to carry out final site clearance; loss of records and information; taking up land that could be used for other purposes; uncertainty over future economic circumstances and regulatory standards; and even unpredictable developments arising from events such as financial crises, pandemics or wars.

Increasingly it seems NDA regards the substantial cost of preparing the site and of ongoing ‘care and maintenance’ would be better invested in early dismantling. Also built into NDA’s decision process are changes in the financial background. Deferred decommissioning assumed that there would be a substantial cost reduction over time (on a discounted or Net Present Value basis) but we are in a period of low or zero inflation that seems likely to persist. And the NDA is clearly anticipating battles ahead in maintaining funding for its activities. ‘This may justify a continuous decommissioning strategy that incurs cost now to avoid unproductive maintenance at a later date,’ it says in the document.

In response, site operator Magnox Ltd is reconsidering the decommissioning programme for the ten Magnox sites. A new strategy emerging would see the company identifying the best candidates for early dismantling – those where the NDA could get most value from releasing land for other activities, and those where the dismantling process would provide ‘the greatest learning for other sites’. Meanwhile, in the overall programme, the emphasis would be on scheduling a steady pipeline of dismantling projects that would retain skills, knowledge and staff, rather than incurring the cost of placing the entire fleet into ‘quiescence’ and then trying to revive the skills base to dismantle the fleet at some time in the future.

Work on the Magnox fleet

In May 2015 the NDA announced a new organisational structure for Magnox Ltd, which runs the Magnox sites on the NDA’s behalf. NDA said the new structure would ensure Magnox Ltd is appropriately resourced to deliver its decommissioning programme. The authority said staff numbers at the 12 Magnox sites had diminished for a number of years as progress was made on decommissioning and the announcement laid out plans to reduce workforce numbers further.

In the new structure all of the common decommissioning projects across the Magnox sites have been grouped into strategic programmes. Although every site is different, the core challenges are similar. Wylfa is different from the other sites, however, in that it has a different plant design and no ponds, which usually contain fuel element debris (FED) and produce resin wastes.

Across the sites Magnox Ltd has four work programmes covering: ponds; intermediate level waste management; FED; and plant and structures. The operator has also created, along with Magnox Waste Management division, a Magnox Care and Maintenance Hub to provide what Magnox has described as a "one-stop" decommissioning service for all the sites.

“Working this way allows us to use innovative techniques and methods as well as providing a sequenced approach to work using the ‘lead and learn’ approach," according to Magnox Ltd.

The principles of the Magnox Ltd strategic programmes are: consistent solutions; mobile teams that integrate with sites, allowing hands-on learning; single points of accountability; joined-up supply chain solutions; affordability; improved flexibility; and sites that are managed in an improved sequence.

Cavendish Fluor Partnership, which was appointed as parent body for Magnox Ltd and a second company focused on decommissioning R&D facilities in 2014, said the ‘lead and learn’ approach delivers locally while integrating the process and programme across all the sites, increasing knowledge, reducing risk and improving value for money. It is also exploring standardisation, for example modularising safety and environmental cases using a special software tool and adapting it to provide a site-specific solution.

Wylfa is scheduled to undergo defueling – expected to take around three years – followed by preparations for care and maintenance. Wyfla’s defueling programme will result in 99.99% of radioactive material being removed from the site.

Each of Wyfla’s two reactors has a large graphite-moderated reactor core. Each core and its single associated steam generator sits within a pre-stressed concrete pressure vessel. Fuel elements removed from the reactors are cooled onsite in dry cells – the only time this approach was used in the Magnox fleet.

There are currently around 88,000 fuel elements at the site which will to be transported to Sellafield for reprocessing. During the defueling period, other hazards will be addressed such as the CO2 coolant, bulk chemicals and oils.

Unlike other Magnox sites with spent fuel ponds, spent fuel at Wylfa is placed in a dry store which consists of a vessel with a series of closed standpipes into which the fuel is loaded from the reactor. The fuel in the dry store cell was cooled with CO2 during operations. Wylfa has three primary dry stores and two secondary stores.

Storing fuel in the Wylfa dry store during operations meant that, unlike other Magnox reactor sites, there was no need to discharge any significant radioactive liquor and consequently there was no need to use a significant amount of chemical resins to filter out radioactivity. It also means there were no chemical reactions occurring between fuel and water leading to contamination of concrete surfaces.

But there have been occasional problems with the Wylfa dry store. They include damage to a few fuel elements by water that entered dry cell four through its roof. The water, interacting with Magnox fuel, damaged some of the elements. Work is now underway to complete the removal of the last pieces of the damaged elements.

Fuel can be removed for transport from either the primary cells, after a minimum cooling time of around 90 days, or from the secondary cells.

On the current timescale, set before the review discussed above, decommissioning work at Wylfa will begin after the defueling phase is complete and the plant is scheduled to enter the care and maintenance phase in 2026. Final site clearance would begin in 2096 and be completed in 2105.

The timescale to care and maintenance is shorter than at any other Magnox plant and is due to the different plant design at Wylfa as well as the absence of FED and resins associated with the use of fuel ponds. FED mainly consists of parts of the magnesium alloy fuel cladding which, at some sites, was removed before the used fuel was sent to Sellafield. More FED remains at other Magnox sites and requires retrieval, treatment and storage.

The Magnox intermediate level waste (ILW) management programme includes retrieval and packaging of operational ILW for final disposal. The total volume of ILW is approximately 6,400m3. The majority of the waste is currently stored in underground vaults and it must continue to be stored onsite because there is not yet a final nuclear waste facility in the UK.

At most sites, long-term onsite storage will take place after the ILW is retrieved and placed in "MiniStores". The concept is that the ILW boxes are used to store the waste until a geological disposal facility is developed and, ideally, are then transferred to the repository. But their introduction as a type of waste package for a final repository in the UK would require assessment and approval by regulators.

Magnox has adopted a generic design for the MiniStores in order to reduce costs. The design allows interim storage facilities to be designed and constructed to accommodate different numbers of ILW packages, depending on need, while retaining aspects of the proven design.

Planning and preparation

Decisions on nuclear waste treatment are critical when decommissioning a site and planning strategically across a number of nuclear sites can result in significant cost savings. Magnox has a Waste Management division, responsible for maintaining waste inventories, developing disposal routes and Magnox’s overall decommissioning strategy, including engaging with regulators.

The Waste Management division provides an interface between waste inventories at the various sites, the waste disposal routes that are open to Magnox, site end states and the projects that link all these constituent parts together. All waste streams are managed through a single Magnox waste management strategy which aims to ensure a consistent best practice approach.

Magnox Limited also has a Plant and Structures Programme. This includes removing plant and demolition and remediation of structures, buildings and land before sites are left in a passive state. The removal of hazardous material, such as asbestos, is carried out as part of this programme. Construction and preparation of facilities that will be required during the care and maintenance period also takes place.

The Magnox Care and Maintenance Hub has been set up to manage sites from their entry into a care and maintenance state through to final site clearance. Its primary aim is to ensure each site remains safe and secure and that all waste stores are emptied before final site clearance.

A generic care and maintenance specification has been laid out for all sites to enable a single set of simple, fit-for-purpose arrangements to be applied during care and maintenance. During this period each site will be passively safe as far as reasonably practicable, with a number of safestores formerly the reactor buildings, as well as a waste store plus any additional buildings that are required. They will be monitored and undergo by planned maintenance and inspection activities.

Fuel management decisions

One of the most important decisions in the Magnox dismantling programme is the fate of spent fuel removed from the reactor. The NDA’s five-year business plan for 2016-2021 says defueling will start at Wylfa in 2016/17.

The foundation of decisions on what should happen to Magnox fuel is the NDA’s ‘Magnox Operating Plan’ (MOP). This document outlines NDA’s strategy for Magnox fuel: to discharge the spent fuel from reactors and then transport it to Sellafield for reprocessing, conditioning and storage of the reprocessing products.

The scope of MOP is management of new and spent Magnox fuel on Magnox sites, including the generation interface and interim storage, transport of fuel from Magnox sites to Sellafield fuel handling plant (FHP), transport of spent fuel from Calder Hall to FHP, transport of Dounreay Fast Reactor material to FHP, interim storage at Sellafield and decanning and reprocessing. MOP also includes contingency plans that aim to ensure Magnox fuel is handled and stored in a safe manner in the event of interruptions to the programme.

The major uncertainty in this plan is the reliability and lifetime of the Magnox fuel reprocessing facilities at Sellafield. The NDA’s ninth MOP, published in 2012, acknowledges this risk. In fact the main difference between MOP 9 and its predecessor is ‘explicit recognition of the uncertainties associated with Magnox reprocessing, primarily the age of the plants involved, which has led to variable delivery performance’, NDA says.

MOP 9 uses a performance range rather than a single delivery schedule and NDA admits this reflects ‘real uncertainty in delivery’.
In January 2014, they published Magnox Fuel Strategy Contingency Options, an analysis of what could happen in the event of ‘unexpected and irreversible failure’ of the Magnox reprocessing capability, with a strategic position on Magnox contingency arrangements.

It examined a range of credible reprocessing failure scenarios, the quantity and location of unreprocessed fuel at the point of failure, plus the circumstances in which contingency plans can be deployed.

It focused on developing fuel drying and subsequent dry storage technology; the case for extended interim wet storage; and the case for extended in-reactor storage. The NDA said the three options were ‘all are considered feasible options capable of contributing to the safe interim management of spent Magnox fuel should reprocessing capability be irreversibly lost’.

The development of drying technology to manage the wet portion of the spent fuel inventory – such as from other Magnox reactors which, unlike Wylfa, stored spent fuel in pools – is at an advanced stage,according to the NDA and there is ‘high confidence’ that this option is deployable. There is also ‘high confidence’ the safety case for extended in-reactor storage can be made and implemented.

‘Extended interim wet storage, by comparison, is less mature, in line with the expectation that this option could serve as an enabling technology,’ it states.

The NDA said that for each option further effort is required: for fuel drying the deployment timescales of the vacuum drying option had to be accelerated; for wet storage the extended timeframe had yet to be qualified; and for in-reactor storage management requirements had yet to be formalised.

In its draft strategy published in September 2015 – 20 months after the Fuel Strategy Contingency Options – the NDA clearly still had concerns over the fate of spent fuel. It reiterated that ‘In the next five years we expect that the Thermal Oxide Reprocessing Plant and Magnox reprocessing plants will complete their committed reprocessing programmes. But it added, ‘There are risks with both Magnox and oxide reprocessing that mean it may simply not be possible to reprocess all of the fuels that are currently scheduled to be reprocessed. We will, therefore, continue to invest in developing alternative options and contingency plans in the event that our reprocessing and storage facilities cannot fulfil their current commitments, or are not available.’

In the current five-year programme, published in February, NDA says its total planned expenditure for 2016/2017 is £3.2 billion, of which £2.25 billion will be funded by UK Government and £900 million by income from commercial operations. Of this, the Magnox programme will receive £550 million and Sellafield £1.17 billion.

In light of the NDA’s work on alternatives to reprocessing Magnox fuel, and its review of the care and maintenance strategy, it appears that the NDA is taking a more pragmatic view of how the UK decommissioning legacy can be dealt with in shorter timescales.

Recognising and acting on the uncertainties of reprocessing is a realistic approach that, when combined with the benefits of maintaining skills and knowledge of the Magnox fleet for timely decommissioning, could prove to be critical for a successful decommissioning strategy for the UK.