Sand in the works21 January 2011
In April 2010, Nuvia Limited completed a project to stabilize about 300m3 of radioactive sludge on the former UKAEA Winfrith site in the UK. Although successful, the project was hampered by the discovery of unexpected materials within the storage tanks. By Madoc Hagan, Rowland Cornell and Keith Miller
At the former United Kingdom Atomic Energy Authority (UKAEA) research site at Winfrith in Dorset, England, Nuvia Limited was contracted to design, build, commission and operate a waste treatment plant to stabilize 300m3 of radioactive sludge stored in four concrete storage tanks – The External Active Sludge Tanks (EAST). The sludge was generated during the operational lifetime of the Steam Generating Heavy Water Reactor (SGHWR), which was shut down in 1990 and is currently in the early stages of decommissioning. In recent times UKAEA has been reorganized and responsibility for the site now lies with Research Sites Restoration Limited (RSRL) with funding provided by the Nuclear Decommissioning Authority (NDA).
The project started in 2000 and was essentially completed in April 2010 (although unexpected sludge, sand and gravel was not removed from the tanks until mid-June). Much of the intervening time was spent in design and building the special plant (referred to as WETP – Winfrith EAST treatment plant) to undertake the operations. Active operations commenced in 2007 and since then several articles have been written describing the plant and plant processes [1, 2]. These references allow this article to concentrate on the aspects associated with the discovery of unexpected materials in the tanks and how they were recovered and disposed of.
EAST was constructed to accept radioactive effluent and exhausted ion exchange resins from the operation of SGHWR. Each tank measures 7.3m x 7.3m x 4.3m internally and has 0.6m thick walls. They were subsequently fitted with a set of four submersible stirrers mounted from the roof structure to provide a means of suspending and homogenizing the sludge.
Although of relatively low specific activity (~35kBq/g), the sludge was unacceptable for disposal at the Low Level Waste Repository (LLWR) site in Cumbria due mainly to its high carbon-14 content. It therefore had to be stabilized and then stored on-site until a time when disposal options will be reviewed.
The WETP facility was built to receive and stabilize the SGHWR sludge into 500-litre stainless steel drums. The design and manufacture of the drums formed part of the contact between Nuvia and RSRL. Below-ground pipework was used to transfer the sludge and liquors from the EAST to the treatment plant.
The sludge had to be homogenized by stirring within the tanks and sampling to demonstrate that it contained the required solid-to-liquid ratio (~26%) for stabilization in the 500-litre drums. This was to meet the disposal requirements using an approved formulation of three ingredients; sludge, OPC (ordinary portland cement) and BFS (blast furnace slag). This had been developed by RSRL ahead of the contract award and approved for use by the UK’s regulatory authorities. The solid-to-liquid ratio was adjusted after stirring by removal of calculated quantities of supernatant water.
The emptying of sludge from the four tanks was completed in early 2010, producing a total of 1068 drums that have been placed into the treated radwaste store (TRS) on the Winfrith site. The latter stages of the sludge stabilization saw the introduction of a separation and filtration vessel (SFV) into tank 1 to hold and then homogenize the residual sludge (termed the ‘final heel’) from all four tanks; we will explain how this was undertaken later.
Sludge was homogenized using the stirrers and, after removal of excess water, pumped across in batches to the WETP plant where it was held in a storage vessel and kept stirred to supply ~350-litre batches to each drum. The OPC and BFS mixture was then added to the drum which had a sacrificial paddle to provide for full mixing of the contents for a specified period. After curing the drums were capped off, lidded and swabbed to check for contamination ahead of removal to the store.
The only access into each tank was via four small rectangular openings cast into the roof structure. The sludge included a significant quantity of radioactive ion-exchange resins and other radioactive particulate debris. Its mixed radiological fingerprint contained about 60% Cs-137. During the sludge homogenization of the full tanks this gave significant doses rates in the 1-2mSv/h range at these openings. For this reason, close attention had to be paid to dose minimisation throughout operations.
The tank emptying process comprised three phases. It started with the removal of bulk sludge from each tank down to around 400mm from the base using a peristaltic pump mounted in the central roof area. (The electric stirrers could not be used below a 400mm depth due to their design). The stirrers were run for about 24 hours to homogenize the sludge and a sample was then taken to determine the solid to liquid ratio using weighing and evaporation techniques. After removal of excess supernate, many batches of sludge were passed to the WETP facility and stabilized into 500-litre drums, a process that mostly proceeded without incident.
The second stage concerned the consolidation of the residual contents of each tank (materials below 400mm depth are termed ‘bulk heel’) into only one tank. At this point the composition of the sludge was secondary to getting it all into one tank where its greater depth would allow the original stage of the operation already described to be repeated. A small submersible pump lowered into each tank via a roof port transferred material into tank 3.
Once the sludge reached a level of around 25mm from the base of the tank, a high-pressure water jet was used to help move the sludge toward the submersible pump for removal.
During the latter stages of the retrieval of the bulk heel from tanks 1 and 4, it was discovered that there was a quantity of what appeared to be a gravelly material greater than 2mm in diameter together with some bolts and other extraneous items (Figure 1) in the base of those tanks. The origin of the gravelly material is unknown but may have entered the tanks during construction of the roof or the later addition of extra concrete shielding. The presence of materials greater than 2mm in size was significant because the 500-litre drums were not legally permitted to contain any materials above this size during the stabilization process.
Various methods were used to recover the metallic items from the tank, including using magnets on cords and other ad-hoc processes. Recovery of the gravelly materials was more problematic. Larger chunks were recovered using a dustpan-type unit; long poles pushed the items into it, and then it was lifted to the roof. In order to assist with the recovery of the remainder, a Big Brute industrial vacuum pumping system (Figure 2) was purchased and used to transfer these materials between the various tanks.
This powerful vacuum system was needed to lift the heavy debris approximately 4.5m from the tank base for transferral between tanks. Tests showed that the equipment was ideally suited for this process and it was extremely successful in operation.
After fitting a recessed 2mm filter screen into the tank 2 roof port, the Big Brute was used to transfer small batches of sludge and associated debris from tank 3 into tank 2 so that the oversize materials could be intercepted, rinsed and removed with the Big Brute. This ensured that the resulting sludge contained only solids less than 2mm in size, in accordance with the terms of the letter of compliance required for production of the resulting 500-litre drums.
Some 1700kg of gravelly materials were recovered from the tanks and placed in batches of approximately 25kg into 68 prepared 200-litre LLW drums. The batches were recovered from the 2mm filter located over tank 2 into the prepared can located outside the tanks at ground level.
The approximately 60m3 of bulk heel in tank 2 was then processed as before down to the last 400mm depth, when the stirrers became ineffective.
When about 20m3 of consolidated sludge was left in tank 2, a further problem was encountered with the discovery of a quantity of approximately 3m3 of sandy residues in the bottom of the tank. These small-grained materials had clearly passed through the 2mm sieve and tests showed that they contained high levels of silica.
The radioactivity of this debris was about as significant as the main sludge batches, with a dose rate of approximately 250µSv/h from a 100cm3 sample. At this point the final phase of the recovery and processing began.
The requirement for another vessel to hold the final heel had been planned much earlier since a means had to be found to ensure that this material met the target specification. Thus a 1.5m diameter, 4.7m high steel separation and filtration vessel (SFV) was constructed with support legs to hold batches of about 6m3 of sludge and supernate. The base was domed to accommodate a stirring system and an outlet pipe at the bottom for the contents when ready for stabilization.
The SFV was lowered through an enlarged opening in the roof port of tank 1. The port had previously been stitch-drilled and sufficient concrete removed to admit the tank.
The Big Brute was used to move most of the final heel into the SFV, except for a quantity of sandy materials which remained in the tank 2 base for later attention. The final heel was then stirred to homogenize it and, after sampling, excess water was removed to allow the final batches of sludge to be pumped to the WETP facility for encapsulation into 500-litre drums.
This process did not go as smoothly as had been hoped; during pumping of sludge to WETP, the pipe-work became blocked with sandy materials, (Figure 3). After due assessment of the situation, it was decided to lower a new suction pipe from the top of the vessel down to a point above the base to minimise sucking up the sandy materials collected at the domed base. After some adjustment this worked well.
The recovery of the sandy material from the blocked pipe-work was problematic, but most was rodded, agitated, sucked or teased out, with only about 1m of the pipe-work close to the base left for later consideration. (The blockage and will be cleared or removed during final decommissioning operations.)
The completion of the encapsulation of the SGHWR sludge in the final heel then allowed attention to be focused upon the handling for disposal of the approximately 3m3 of sandy materials.
The Big Brute recovered the sandy materials from tank 2 and the SFV into a disposal container located in the centre of a standard 200-litre LLW drum. Surrounding the disposal container inside the drum as shielding was crushed contaminated demolition waste. Cement and water were then added and the contents were stirred with a small remote-controlled machine to homogenize them and allow them to stabilize when the mixture cured.
In this situation, dose rates were measured dynamically on the outside of the LLW drum to ensure that the local acceptance level for LLW disposal was not exceeded. Trial and error showed that between 15 and 20kg of sandy material was appropriate. The drum was then covered and moved to a shielded location to cure; then it was capped off with more crushed concrete and the lid attached. A total of 186 of these drums were produced during this process, which stabilized about 3m3 or about 2.75Te of this debris. The Big Brute remained at the roof level and away from operatives at all times to minimise dose uptake and local shielding was also used to assist in the achievement of this objective.
The project methods were developed on the basis of historical information supplied by the client. This information included details of the tank contents, much of which was based upon relatively recent homogenization and sampling campaigns. A clear lesson is to be more challenging with regard to the methodologies adopted to provide historic data. In this instance, had the probability of oversized material been recognised through discussions with the client, it may have been possible to sample the tank bases for these materials during the design stage.
Another inevitable problem associated with these unexpected items was the need to update safety and operational documentation in a timely manner, and to minimise the delay in obtaining approval for these documents from the client. At times this interface could have run more smoothly in order to proceed with operations.
During this project there was considerable attention to operator dose minimisation, but the presence of the oversize materials and latterly the sandy debris significantly added to this burden. Thus the collective doses during the recovery and stabilization of these two sets of debris were 5.3mSv and 7.5mSv, respectively. Despite these exposures within a team of about eight operatives, no Dose Restraint Objectives (DROs) for the EAST /WETP plants were exceeded. (The DRO for 2009-2010 was set to be 4mSv/y per operative. In 2011 this is to be reduced to 2.5mSv/y)
Crucially, the presence of gravelly and sandy debris at the bottom of each tank was seriously challenging to the timescale, and thus the costs, of the project. It has been estimated that the task of recovery, stabilization and disposal of these materials added some months to the project duration. It also produced 254 additional LLW drums for disposal.
Overall the processes devised to handle the main bulk of sludge worked well. The plant reliability was generally excellent and the Big Brute machine worked very well and helped minimise operator dose. The use of the SFV was partially successful, but the intrusion of the sandy debris reduced its efficiency and challenged its original design.
It is instructive to look at this whole project at its completion and surmise how it could have been undertaken differently had the presence of the oversize and sandy materials been known at the outset. Although the planning for this project would have been different, we conclude that the actual methods used were effective. The main difference would have been the provision of recovery and disposal routes ahead of time, rather than having to develop new methods with urgency when the needs arose.
Clearly the SFV design would have been changed to perhaps provide a more vigorous stirring capability and clearly the withdrawal pipe-work should not have been located at its base. The degree of difficulty in recovery of sludge from the base of each tank was not fully appreciated at the outset but the standard industrial equipment purchased was fully satisfactory and upon reflection would have formed part of a pre-planned operation.
The problems noted above demonstrate the difficulties often experienced with projects of this nature on old plant and facilities where unexpected problems nearly always emerge in the course of the work.
The authors are Madoc Hagan, Rowland Cornell and Keith Miller from Nuvia Limited, both located at the Winfrith Site in the UK.
 Hall, C., Hagan, M., Cornell, R.,