Thinking outside the cage25 March 2010
A concrete backfill with a water barrier has been developed to stabilise legacy waste packages stored in the 1960s-70s at the Richard radioactive waste repository in a former Czech limestone mine. The barrier, a so-called hydraulic cage, provides an alternative pathway for groundwater to prevent the build up of a pressure gradient across the concreted waste body, and can be used even in saturated rock. The design could be applied to other waste repositories worldwide. By Bernt Haverkamp, Miroslav Kucerka and Enrique Biurrun
The Richard repository for radioactive waste from institutional waste producers is located in a former limestone mine, near Litomerice in northern Bohemia, the Czech Republic. It has been in operation since 1964, accepting low-level waste from hospitals, industry and from research involving radioactive materials.
In 2000 the Czech Radioactive Waste Repository Authority (RAWRA) took over responsibility for radioactive waste management, disposal and operation of the country’s three repositories, including Richard.
Following the takeover it came to light that there was a lack of detailed information on the estimated 20,000 packages of historical waste disposed of at Richard in the 1960s and 1970s. Therefore, RAWRA began preparation of a project to inspect, relocate and stabilize the waste packages.
The project was supported by the European Union through the Phare programme, which provided assistance to new EU countries. Approximately EUR1 million in funding for the project was came from the EU while the Czech Republic provided EUR450,000.
The Richard mine was excavated into an almost-horizontal, 4m-thick layer of limestone embedded into two layers of marlstone. The elevation of the site is about 265m above sea level. Towards the top, the repository caverns are covered by 30-70m of marlstone. The marlstone layer below the repository is approximately 50m thick and is followed by a layer of Cretaceous sandstone with a thickness of around 100m (Figure 1).
The sandstone layer is a typical aquifer, from which water is pumped up to supply Litomerice and the surrounding area. The aquifer in the sandstone is fed by infiltration from above. Data provided on the location of the water table are not consistent, but the water table is some tens of meters below the repository and could be located in the marl or sandstone. Thus, the repository is situated in partially saturated rock.
The complete mining system called Richard is an extensive network of tunnels and caverns and is subdivided into three distinct areas: Richard I, Richard II and Richard III. About half of the existing cavities in the Richard II mine are used as a repository. Richard II is accessed through a 100m-long horizontal tunnel with an entrance on the slopes of Bídnice Hill above Litomerice.
A layout of the central part of the mine at the start of the project is shown in Figure 2. The chambers in use for waste disposal are shown in pink, the access tunnels are coloured light brown and the non-operated parts in yellow. The chamber-system 8/2-9-12 designated for closure within the course of the Phare projects is indicated in the lower left part of the figure.
The waste disposed of in the repository is mainly a mixture of low-level radioactive waste (LLW) and short-lived intermediate-level radioactive waste (ILW), although the wastes do contain some long-lived radionuclides (americium-241, plutonium-238 and -239). The wastes consist of solid material, low activity liquid wastes and sludges. A significant inventory of organic material may be present (including paper, wood, rubber, textiles, plant wastes, bedding, straw, and animal carcasses). Up to 25,000 waste packages with an activity of 1015 Bq have been disposed of.
The waste is usually stored in 200-litre drums, which have a 100-litre drum cemented into them so that every standard waste package has a radiation shielding of about 5cm of concrete. However, during the early years of repository operation, from 1965 to 1984, a significant amount of waste was disposed of in 50-litre drums of galvanized steel.
In compliance with respective regulations in the Czech Atomic Act, RAWRA developed a preliminary plan for the closure of the Richard repository and a safety assessment (SA) demonstrating the long-term safety of the disposal facility  as part of the repository license documentation.
The closure concept, which formed the basis for , detailed plans to backfill the waste chambers with concrete. The filled waste chambers were to be sealed by concrete walls, through which high-quality concrete was to be pumped to completely backfill the chambers. The concrete for backfilling was supposed to have a low hydraulic conductivity, below 10-10 m/s, and low shrinkage. The rest of the repository (main drift, entrance, currently used ventilation shaft etc.) was to be sealed by concrete plugs. Adjacent tunnels and caverns were to be backfilled with lower-quality material as the only purpose of this backfilling was to ensure the geomechanical long-term stability of the repository.
Well in advance to the closure of the repository, the concept recommended the sealing of the former ventilation shafts, which are supposed to be major contributors to the inflow of water into the existing drainage system of the mine.
However, very little information is available about the rate and nature of water infiltration in the area of the repository. The soil, marlstone and limestone are believed to have relatively low permeabilities. It is not known if water would infiltrate fairly uniformly over the area of the repository once the repository is closed, or if it would flow through the repository heterogeneously, as the area exhibits a number of faults. For  it was assumed that a certain percentage of the precipitation infiltrates the marlstone in the overburden, percolates downwards and through the repository, advectively transporting radionuclides out of the former mine.
The source term (the types, quantities, and chemical forms of the radionuclides that encompass the source of potential for exposure to radioactivity) in  is modelled as a homogeneous mixture of waste and concrete filling the parts of the repository currently in operation.
Two release scenarios (apart from human intrusion) have been identified. For both the town well scenario and the farm scenario, it is assumed that the contaminated water from the mine travels downwards through the lower marlstone into the aquifer, from which – at a certain horizontal distance – contaminated water is pumped up again to serve as drinking water in the first scenario or as water supply for the operation of a small farm in the second scenario.
Based on the assumptions on future repository development and within the limits of the used models, for both scenarios  rendered exposure values for future generations, which are below current regulatory limits.
As a third potential route of radionuclide release, the direct release of contaminated mine water into the biosphere was considered, caused by bypassing the access tunnel sealing.
To assess the worst possible exposure connected with such a direct release, a reference case was made up. For this reference case it was assumed that water percolating into backfilled chambers becomes contaminated with radionuclides, is swiftly discharged out of the former repository system without any dilution, and becomes the sole source of drinking water for members of the critical group. Such an extreme worst-case scenario, not surprisingly, would result in radiation exposures well above acceptable values.
Notwithstanding the radiological risk associated with this reference case, it was excluded from further analysis in  as a very low probability event on the basis that high performance sealing of the access tunnels would reliably prevent this scenario to occur.
In 2003 RAWRA launched a call for tenders for the review and further development of the repository closure concept and detailed design of the closure chamber in the Richard repository. The contract for this project was awarded to DBE Technology GmbH.
When reviewing the safety assessment, DBE considered that although the rationales of scenario selection and of some key assumptions in some aspects lacked a clear, defensible justification, for the town well scenario and the farm scenario the safety assessment seemed reasonable and their results reliable.
With regard to the reference case, however, it was concluded that the direct release of contaminated mine water could not simply be excluded just by sealing the entrance tunnel in such a way that mine water would not be able to bypass this barrier. Instead, it was judged that for several reasons it would be very difficult to effectively prevent any direct release and even more difficult to prove that such release would be prevented in the long term.
The main reason for this conclusion was that even if total long-term sealing of all access tunnels to the Richard repository could be achieved, which would be difficult to prove, this would still not prevent the possible direct release of mine water. First, the hydraulic conductivity of the underlying marlstone is lower than the hydraulic conductivity of the 3-4m thick limestone layer surrounding the repository, and second it has to be expected that preferential pathways exist along the limestone base or fractures inside the limestone with even higher permeabilities than the limestone itself.
Therefore the probability is considered rather high that water inside the repository would be released through the limestone rather than through the underlying marlstone. Only if vertical fractures with high permeabilities existed in the lower marlstone layer, which could easily be reached from all areas of the mine, would water from the mine be released preferentially through the underlying marlstone into the aquifer.
HYDRAULIC CAGE CONCEPT
Thus, it was considered necessary to develop a changed closure concept, which would prevent the radiological consequences from possible scenarios associated with a direct release of mine water. These considerations lead to the development of a closure concept involving the installation of a hydraulic cage around the waste chambers.
The main feature of the proposed new closure concept is the installation of a hydraulic cage around each disposal chamber to prevent the build up of a pressure gradient across the concrete body, which fills the disposal chamber after backfilling. Without such a pressure gradient there is no driving force for groundwater flow through the concreted waste body and consequently no advective transport of radionuclides out of the disposal chamber.
While the former radionuclide isolation system of the repository was based only on the principle of radionuclide containment by enclosing the waste with low-permeability barriers, the hydraulic cage provides an additional engineered barrier.
The main safety advantage of the hydraulic layer is the prevention of advective flow through the enclosed waste/concrete body. Without the driving force of a pressure gradient, no advective transport of radionuclides would take place, even if due to capillary forces the concrete body soaks up water until it is 100% saturated. This would also be the case if the whole repository system were 100% saturated.
A hydraulic cage is realized by building a highly permeable layer around the chamber as preferential pathway for groundwater that might be present. A simplified cross section of a waste chamber backfilled according to the hydraulic cage concept is shown in Figure 3: a layer of pure concrete surrounded by a high-permeability layer encloses the repository space of stacked waste containers backfilled with concrete.
The normal evolution scenario will thus be changed in such a way that no release of radionuclides will occur apart from diffusive fluxes between the waste/concrete body and the gravel layer. It has to be noted that diffusive fluxes would take place only if the hydraulic layer is filled at least partly with water, that is, if the repository is saturated to a certain extent, which is not necessarily the case. Without such saturation there would be no release of radionuclides at all.
Prior to implementation of the concept an update of  taking into account the changed source term was carried out. The model for assessing the hydraulic cage system performance was developed in close analogy to the model in  using the computer code GoldSim. Changes were implemented only in the source term part of the model in order to quantify the effects on the release of radionuclides caused by the different closure concepts. As it can be supposed that a high permeability difference between the gravel layer and the waste/concrete body will persist throughout the period for safety calculations, for the hydraulic cage model transport of radionuclides is restricted to diffusive fluxes.
The models show that the implementation of the hydraulic cage effectively reduces radionuclide transport during the first thousand years due to elimination of advective flows and low diffusivities of the concrete layer and the concreted waste body (Figure 4). With progressing degradation of the waste/concrete mixture and the pure concrete, diffusivities will increase leading to rising radionuclide transport and accordingly to higher annual dose rates. As time goes on, calculated values for the hydraulic cage model are higher than respective values for the former closure concept model in which most of the radionuclides would have already been flushed out of the repository. The peak value of the annual dose rate is reduced by three to four orders of magnitude and appears now at about 6000 years after closure of the repository.
For the two more likely scenarios, the town well and the farm scenario the changed closure concept also leads to significant improvements in regard to possible radiation exposure. The peak annual dose rates are reduced by a factor of four and the respective peak time is shifted several thousand years into the future. Differences between the closure concepts are less pronounced in these scenarios mainly because differences in the amount of short-lived radionuclides being released from the repository are reduced by natural decay during their migration through the geosphere.
Also, in addition to reducing the potential hazards related to the different scenarios, the implementation of the hydraulic cage concept drastically reduces the probability that either of these scenarios will ever occur at all. For this reason, based on the results of the updated safety assessment discussed above and consultations with Czech authorities RAWRA decided to implement the hydraulic cage concept for the detailed technical planning of the closure of a certain chamber-system in the Richard repository.
BUILDING THE CAGE
Phase 1 – design
The project’s kick off meeting was held on 28 November 2005 and project activities started on 5 December.
The first activities were focused on the development of a detailed design for chamber reconstruction and the preparation of all the necessary documentation for the licensing of planned actions. With respect to the project, there are two main licensing authorities involved – the Czech Mining Office and the State Office for Nuclear Safety. The Phare contractor was Czech firm Erebos. Project subcontractors included Czech firms Tubes (for design), Age (for technical supervision) and DBE Technology (consultancy).
In parallel to the development of documentation, Erebos built its infrastructure for the project realization and RAWRA organized the training of the Erebos personnel on emergency preparedness, radiation protection.
Phase 2 - preparatory work
Full-scale activities concerning chamber reconstruction started on 2 January 2006. First, chambers 4, 5, 7, 8/1 and 10 were prepared and stabilized for accepting debris from the reconstructed chambers. During the clean up, more than 600m3 of debris rock was removed from chambers 8/2, 9 and 12 and disposed of in the aforementioned chambers. This phase was completed in early March 2006.
Phase 3 - construction
This phase included the construction of the hydraulic cage, supporting frames, floor with drainage system and concrete walls. The construction started in the second half of March and the installation of the hydraulic cage on chamber walls and ceiling was completed in June 2006. In July 2006, three segments were completed and prepared for accepting the first waste packages. The follow up construction of the remaining segment walls and bottom were carried out simultaneously with waste package relocation.
Phase 4 - relocation of waste
Phase 4 involved the removal, inspection, conditioning and relocation of historical waste to the new chambers. Waste contractor ALLDECO.CZ, now part of AMEC erected a separate working area in chamber 17 for the inspection and compaction of historical waste packages removed from neighbouring chamber 22.
Each waste package was measured for gamma exposure rate and neutron flux and the contents of each drum were visually checked. Depending on the inspection results, the shift foreman decided to compact the waste package or treat it in other ways. Some of the packages were in very bad condition (mostly heavily corroded). These packages were usually inserted into plastic bags or completely repacked.
Special attention was paid to the radiation protection of the 21 workers and the workplace was monitored for potential biological contamination (bacteria, fungi, spores, etc.). The maximum cumulative individual effective dose during the package relocation period (13 months) reached 8.39mSv and the total collective dose was 87.14mSv. The average monthly individual effective dose was 0.25mSv/month.
Phase 5 - backfilling
For the waste backfilling a special concrete was developed to ensure the filling of all voids between the waste packages. Design requirements of the concrete strength were minimum 37MPa, while achieved values after 56 days of setting reached 55-65MPa.
The chambers were backfilled in 3-4 meter long modules. In the first modules a barrier was created from 200-litre waste drums supporting a pile of compacted and non-compacted waste packages, up to the height of the concrete wall. Then a formwork was erected and the waste pile was backfilled with backfilling concrete. On top of the backfilled waste module the waste packages were inserted in a manner so that a void space between the hydraulic cage surface and the waste packages pile was a minimum of 30cm, to ensure a sufficient isolating concrete layer. The waste packages were fixed against floating using a KARI mesh that was anchored to the concrete walls. After the stepwise backfilling of the chamber segment by module, the remaining non-backfilled upper layer of waste packages was backfilled at one time, so a compact concrete ‘head’ in the segment was created.
During the project a total of 15,000 waste packages were relocated. It was expected that this would cover the total amount of historical waste. However, in reality a further 8000 historical waste packages still remained in the original chambers. Therefore, RAWRA has since adapted chambers 13 and 22 (total volume 1500m3) for disposal of the remaining historical waste packages. It continues to relocate the packages and backfill concrete around them.
Miroslav Kucerka, Radioactive Waste Repository Authority, Dládná 6, 11000 Prague 1 - Czech Republic. Bernt Haverkamp and Enrique Biurrun, DBE Technology GmbH, Eschenstrasse 55, 31224 Peine, GermanyRelated ArticlesDisposal plans (part 4: low- and intermediate-level waste)
 Chambers, A.V., R. Cummings and B.T. Swift: Performance Assessment of the Richard Repository, Serco Assurance, Final Report SERCO/ERRA-0479 for RAWRA, 2002.