Silo B41 was built in 1952 to take solid waste from Sellafield's two military reactors. Looking for a suitable storage structure, engineers selected designs for a Canadian grain silo - a reinforced concrete box measuring 21m high and 24m long by 8m wide, containing six silos arranged in a line. These were accessed by an overhead corridor.
Waste was tipped into the silo compartments through charge holes from a trolley running on rails the length of the corridor.
Aluminium and magnesium oxide fuel rod cladding, graphite from the reactor cores, chemically active waste, irradiated metals, wood, paper and anything else workers wanted to dispose of was slung into B41.
By 1965 the silo was full to the brim with 1,600t of intermediate level waste, and was shut down.
B41 was put largely out of mind for the next 25 years but by the early 1990s it was becoming clear the facility could not be ignored.
Corrosion of metals mixed with chemical waste had produced hydrogen in the silos. Even before the silo was closed this posed a explosion risk as workers carried out routine maintenance on the facility. Following its closure, cracks began to appear in the walls as natural weathering took its toll. This allowed hydrogen to seep out. It also provided an escape path for radiation.
In 1996, then site owner and operator and BNG's parent British Nuclear Fuels (BNFL) decided the silo would have to be emptied of its contents and demolished.
Preliminary works for the task have just been completed. Efforts will now focus on developing a waste retrieval methodology, says BNG head of silo decommissioning Byron Smith.
First the building had to be made safe to work around.
Cracked and weathered concrete needed to be patched. Next, the silo's outer radiation shield was removed to give access to the silo walls. This was possible because natural radioactive decay had reduced the waste's hazard level, rendering the shield redundant.
To prevent machinery igniting hydrogen, inert argon gas was circulated through the silo to expel it. This was calculated to reduce risk by 70%.
Next, a robotic camera was sent into the silo charge corridor.
Engineers were shocked at what they found, says Smith. Instead of being clean and debris free, there was a large mound of waste at the corridor's far end and a trail of waste along its length, he says.
Operators of the silo had run out of capacity in the first chamber far faster than anticipated and, guessing that waste had jammed in the narrow charge portal, had continued tipping in the hope that the blockage would give way under the weight of material pressing down from above.
A generation later, no records of waste volumes or disposal practices were available. The camera yielded black and white images showing the scale of the unexpected problem in the corridor but gave no clue about what spare capacity, if any, was left in the silo itself.
The best way of clearing the corridor was to force the waste into the silo. BNFL contemplated designing and building a complex robot equipped with cameras and tools to clear the blockage, similar to those being used in the decommissioning of reactors, but it posed significant cost and reliability risks.
As an alternative, one of the engineers suggested the blockage be dislodged using a manually operated pole, says Smith. The simplicity of the idea was seen as a virtue. Placing a lead plate on the corridor roof, with access holes cut above the charge portals, meant workers could tackle the waste with scaffold poles - 'pokey sticks' - while being protected from radiation.
The technique worked, and pairs of workers took backbreaking turns ramming the waste through the charge hole into the half empty chamber.
Ramming was done under guidance from a worker wearing 'virtual reality' goggles from a Playstation 2 computer games kit, connected to the camera inside the corridor.
Clearing of the corridor was completed in December 2003, and the decommissioning team moved to preparation for waste retrieval. With no proper floor within the corridor, a new castinsitu concrete work platform had to be built providing a solid surface for access and mountings above the charge holes into which thick steel plugs could be fitted.
To enable handling of materials, plugs and future waste retrieval equipment, a modular steel portal frame, running the length of the corridor, was constructed. Faced with tight space constraints, the team rehearsed all manoeuvres off site in advance.
Work was carried out in difficult conditions. The argon gas and presence of radioactive particles meant the air was deadly, forcing workers to wear breathing apparatus. The delicate operation of plugging the charge holes was completed in 2004 using a winch and pulley system. And, brick by brick, the corridor's walls were demolished.
Efforts are now focused on designing the equipment and method for extracting the huge volumes of waste that fill the silo.
The first plan was to build a new plant over the silo to remove, process and package the waste.
However, craning this into place would be almost impossible given space restrictions on site.
Engineers were forced back to the drawing board. Target dates for the next phase have yet to be decided.