Hive of activity

Over the past half-century, the eyes of the nuclear industry have often been focused on a remote, windswept corner of Northern England - Windscale, and the surrounding Sellafield complex - and sometimes the attention has been for unfortunate reasons.

But now, the focus is on technology, as engineers from BNFL and UKAEA attempt a feat which nobody has managed - dismantling a large-scale gas-cooled nuclear reactor.

The Windscale Advanced Gas-Cooled Reactor (WAGR), occupying the instantly-familiar golfball on the edge of the Sellafield complex, is now the home for more activity - human activity, that is - than at almost any time in its existence.

Originally built as a working prototype for AGRs, WAGR was shut down for good in 1981, and has now effectively reverted to its original function - an R&D platform. The decomissioning campaign, Project WAGR, aims to prove that a power reactor can be taken apart safely, predictably and cost-effectively, using existing technology.

With over 90 similar reactors around the world currently awaiting decomissioning, the project is providing valuable information - and personnel - for future projects.

It's been necessarily slow going. The reactor was built in the 1950s, and its original designers never considered the possiblity of dismantling it. Decommissioning, handled by main contractor BNFL Environmental Services, began in 1983 with the removal of the reactor fuel, and much time since then has been occupied with preparatory work, such as installating an automated dismantling machine with a remote-controlled manipulator arm and overhead crane. In December 1999, the equipment and tooling was in place, but dismantling had yet to begin.

Things have proceeded rapidly since then. The reactor core has been exposed and is currently being removed, graphite brick by graphite brick. Almost two-thirds of the core has been removed, and the project is now some six months ahead of schedule. Project director Terry Benest expects the current phase of the project - the removal of all the equipment apart from the concrete biological shield surrounding the reactor, and the aluminium-clad steel golfball itself - to be complete by late 2005.

The current good progress is despite - and in some cases because of - several drawbacks on the way. Several times, the reactor has defied expectations. The very first phase of reactor decommissioning, the removal of the manifold which took the hot coolant gases from the reactor into the heat exchangers, known as the hotbox, set the scene for the problems the team would encounter. The complicated steel structure was far more radioactive than had been expected, and some construction details had not been recorded on the original drawings.

The next phase, removal of the six loop tubes running through the reactor core, also proved problematic. These tubes, which were used as experimental fuel channels, were among the most radioactive parts of the plant, running through the centre of the core and receiving high doses of radiation. The engineering team decided that the best way of cutting the tubes was with a hydraulic shear, as this created less dust and vapour than a saw, grinder or cutting torch. However, before the tubes could be cut, they had to be filled with cement grout to prevent them from collapsing as the hydraulic jaws tightened. It was not possible to do this remotely, so engineers clad in protective clothing had to do the work by hand in the 'crypt' below the reactor.

What helped here was working on mock-ups of the equipment. 'The engineers rehearsed the operation, wearing full protective clothing, until they were extremely slick and could complete the task quickly,' explains Benest. Mock-ups have been invaluable to the project, he says, allowing the BNFL Environmental team to develop tools and techniques ahead of time, and helping operatives come to grips with the tricky handling procedures for each stage of the project.

Mock-ups are once again the key to the current activity, the dismantling of the core, which is made up of eight layers of interlocking graphite blocks, held together with steel restraint bands and interlaced with flux scanning tubes and thermocouple wires. The restraint bands have to be cut before any of the blocks can be removed, and this is being done with a reciprocating saw which clamps onto the bands.

The clamp, a two-piece steel structure, was designed in the mock-up stage, and is deployed by the robotic arm. Each block, 80cm long and weighing around 60kg, is removed using a ball grab - a metal post, studded with spring-loaded steel balls, which grip the inside hole running through the centre of the block. The grab, which is moved by an overhead crane, places the blocks inside the metal cages that fit into the concrete waste storage boxes, which are later filled with cement grouting and finally sealed with a layer of concrete.

Core removal is expected to take several more months and will create a large volume of intermediate-level nuclear waste, which is currently being stored in the former turbine hall of the AGR, a concrete-walled building. The boxes are to be stored there for at least 50 years, until a site for a permanent UK nuclear waste repository is chosen and the facility built. It could then be a further 50 years until the waste is transported to the facility.

The next phase in the decommissioning is the removal of the heat-shield around the core. This comprises three two-inch-thick layers of steel, loosely bolted together. These plates should simply lift out of the reactor, but the team has already developed two contingency plans - if they have stuck together, they will be jacked apart; and if this doesn't work, they'll be cut apart with a 0.25MW oxypropane cutting torch.

'We've found that developing contingency plans saves us a lot of time,' Benest says. 'It's good insurance, really - we don't have to stop completely to develop new methodologies and tooling.'

Once the heat shield is out of the way, the team will begin to cut up and remove the lower structures of the reactor, which supported the core, using a remote-controlled 0.25MW oxypropane cutting torch. Then it's the turn of the pressure vessel itself, made from carbon steel, varying from 44mm to 111mm thick, and surrounded by three layers of asbestos-rich cement insulation blocks, up to eight inches thick in total.

The presence of asbestos adds a further safety issue to the existing radiological risk. The team plans to use the oxypropane torch to cut through the vessel, but its 25-foot flame will be augmented by iron powder injection to allow it to cut through all the layers in a single pass.

Benest estimates that the project will be complete, with just the biological shield and the golf ball still standing, by the end of 2005. In all, the project will cost £80million. One of the most complex pieces of reverse engineering in history will be completed on time, and within budget.