Skip to main content

Prototype repository - full scale testing of the KBS-3 concept for high-level radioactive waste


The engineering practice applied on underground construction and operation is by tradition tougher than the practice applied on surface objects. Another difference by principle is that the work-space needed underground has to be created while the space on surface is unlimited. In case of a repository the tolerances and accuracies are the same under and above ground. But enough large space has to be provided underground. The project adopted and applied surface engineering standards with satisfactory results. The quality of the work is controlled in accordance to the principles of ISO 9001 and 14001. The principle was developed to include planning, checking that each activity in the plan is carried out and checking that the described quality in the plan is obtained. This checking was developed into a system of traceable documentation on what was made and how. Independent technical audits were introduced as regular parts of the system. Based on the experience, which is good, a robust system may be formed for which procedures are known and tested, as well as the need of staff.
The report makes final predictions of the chemical changes in the engineered barrier system (EBS) materials and pore water compositions in the Prototype Repository. Considering the geochemical performance of a repository, the central issues are what kind of pore water compositions a repository is expected to generate and how these pore water compositions develop as a function of time. The first wetted volumes of buffer and backfill are created at the boundaries of the Prototype EBS. These first pore waters are sucked deeper into the under-saturated volumes of buffer and backfill and represent the first wetting pore water compositions of the EBS. At the same time, new pore water compositions are repeatedly generated at the EBS boundary. These waters, as well, are likely sucked deeper into EBS until the suction power vanishes to the EBS water saturation. Therefore, the pore water evolution at the EBS boundary and the first wetting pore water evolution in the EBS interior essentially define the limits of compositional pore water gradients that will develop in a repository EBS. The predictive tool built on the PHREEQC-2 modelling software will be calibrated with the laboratory studies available in the literature. The experimental data from the Prototype Repository vessel samplings, if available, will be compared with the modelling and/or the modelling tool will be calibrated to this data as well. The modelling aim to predict pH and pe conditions and Na, Ca, K, Mg, Fe, Si, SO4 and alkalinity concentrations in developed pore waters. In the case of solid phases, the modellings take into account surface complexity and cation exchange in buffer and backfill, and make estimates of mole-transfers at these surface sites. Furthermore, modelling consider calcite, gypsum, pyrite, goethite and quartz as reactive phases that equilibrate with pore waters, and predict concentrations and mole-transfers of these phases as well.
The most significant expected result is a validated model of coupled thermo-hydro-mechanical and geochemical (THMG) processes in a real scale KBS-3 repository type, in which the HLW are disposed in vertical holes, excavated in crystalline host rock. The CODE-BRIGHT numerical code is being used for this purpose, but some new developments and modifications of the numerical model are necessary to take into account the special characteristics of the backfill material and boundary conditions of the problem. There is an important particularity, which is the high salt concentration in the underground water. Therefore the geochemical coupling is very interesting. The model will be constructed in a stepwise strategy. Initially laboratory results have been checked to derive constitutive models. T and TM analyses are then considered to check the mesh and to compare with analyses performed by other groups. The next step is a THM simulation of one canister, focussing on the behaviour of the bentonite. Finally, at the end of the project, a consideration of the chemical aspects involved in the bentonite buffer will be addressed in a preliminary simulation.
The aim is to decrease the size of the deposition tunnel as much as possible, and the smallest dimension requires a deposition method in which the canister is "tilted" into the whole, i.e. by Cardano-movement. This requires a deposition machine which can move forward during simultaneous raising and lowering of the canister. Such a machine has been constructed and tested in two versions: with and without totally surrounding radiation shield. Both machines perform to satisfaction and the stack of blocks in the Prototype Repository had large enough centre openings for allowing emplacement of the canisters as planned.
This work has covered a whole range of numerical analyses to simulate the complicated processes occurring in the Prototype Repository. This has included: a) the simulation of the thermal regime of the full Prototype Repository; b) the hydraulic distribution in the host rock prior to canister and buffer emplacement; c) coupled TH analyses through a canister section; and d) a fully coupled THM analysis through a canister section. The fractured nature of the host rock has also been modelled. The analyses have been conducted in 1, 2 and 3 dimensions for a number of varying domain sizes and under a number of different geological conditions. All simulation work was carried out using COMPASS. Great attention has been given to the material parameters inherent in the Prototype Repository, in particular the buffer and pellets. Hence at all stages of the simulation work it has been endeavoured to validate the numerical results with actual measured experimental conditions. The simulation work aims to model the flow regimes within the system over the 20-year period of the Prototype Repository project. It is envisaged that this work will contribute towards building confidence in the use of engineered barrier systems for the safe and future-proof disposal of high-level nuclear waste.
Two deposition holes will be instrumented with innovative fibre-optics displacement sensors for tracking canister movements during a period of time up to 20 year. These sensors were custom made for withstand the harsh environment into which they are required to work (high temperatures, pressures and corrosive media). Furthermore, the sensor's anchoring system to the rock and to the canister was carefully designed and developed in order to solve all the difficulties regarding the canister emplacement method but keeping the required tracking system functionality. Of course, the good performance of those instruments and anchoring system will depend not only on their adequate selection and construction but also on the installation procedure, given the expected test objectives and test site constraints. The final result will be the evaluation of the whole tracking system design, construction and installation techniques and the sensors reliability and performance in this environment, after a period of up to 20 years. The gained experience and acknowledge could be used to suggest modifications and improvements for future tests and monitoring in real repositories.
The work has been aimed at developing and integrating COMPASS (COde for Modelling PArtially Saturated Soil) into a sophisticated pre and post-processing package. This has allowed full visualization of large 3 dimensional domains and is of particular importance in the simulation of the Prototype Repository. Due to the increasing computational demands of the work, attention has been paid to the implementation of high performance computing techniques including parallel computing approaches and iterative solvers. Analyses which previously had upwards to 3-4 weeks turnaround can now be speedily performed in a matter of days. This has benefited the work greatly with the outcome being a more comprehensive, detailed simulation of the Prototype Repository in the time permitted. The above work is currently ongoing with modifications and new technologies being implemented regularly. With the further evolution of COMPASS the potential to tackle more diverse and varied applications is of paramount importance.
The deposition tunnels are intended to be backfilled with a mixture of crushed rock and bentonite, which is compacted in situ. The overall requirement on backfilling is that the backfilled tunnel shall have a hydraulic conductivity, which is equal to that of the repository rock on average. But as backfilling provides a more qualified result in the centre of the tunnel and less so in the ceiling and wall regions the main problem is to find efficient method and equipment for those regions. By placing the backfill material layer by layer in a slope and use a vibrating plate for compaction the centre core was more compacted than planned. With special devices for pushing material up to the ceiling region and for compacting the zone closest to the ceiling satisfactory densities were obtained. So far, one of four electrode arrays has been installed in section1 to monitor the water uptake in the backfill by geo-electric monitoring of the material resistivity. Dryer zones initially observed near the tunnel wall have disappeared during the first six months of measurement.