The purpose of this study is to estimate the effect of earthquakes upon a groundwater flow around a repositories for high-level radioactive wastes. Estimation of a groundwater flow change before and after an earthquake or a volcanic eruption is one of the issues for a long-term safety assessment of the repositories. However, almost any systematic investigation about the causality between a groundwater flow change and an earthquake or an eruption was not found, and as well no estimation formula has been published. We succeeded to obtain a primitive relationship between a groundwater change and an earthquake in this study.

Our study consists of three stages. First, several survey reports which describe field observation results of groundwater anomalies caused by earthquakes or eruptions have been collected. The necessary data have been read from the literature and systematically arranged. Second, source mechanisms of the corresponding earthquakes were inspected and static displacements at the well positions were calculated by the dislocation theory in the seismology. Third, parametric studies among the parameters of groundwater anomalies and earthquakes were carried out to find a numerical relationship between a couple of them. Then, a preliminary relationship between water table change in a well and static displacement at the well position was found.

We can conclude that temporary change of water table seems to depend on the norm of displacement vector. In this relationship, the maximum value of water table change would be approximately one hundred times of the displacement.

The Swedish Nuclear Power Inspectorate is developing a new methodology for the construction of scenarios for radiological consequence analysis as part of its SITE-94 performance assessment project. SITE-94 involves the incorporation of site specific data from the Äspö site into a performance assessment (PA) of a hypothetical high-level waste repository. This paper describes a systems analysis approach that has been developed based on the concept of organising all the events and processes which need to be taken account of in PA into a ‘process system’ and a much smaller residual group which are used to generate scenarios. The methodology used for developing scenarios, producing calculation cases and addressing the various types of uncertainty involved in PA consequence analysis is described.

Integrated performance assessments of proposed repositories for HLW involve multidisciplinary studies which include identification of possible paths of future evolution of the repository system (scenario analysis) and quantification of the consequences of each scenario using deterministic or probabilistic modelling approaches. The results of such studies are commonly presented as estimates of consequence (e.g. dose to a representative individual) against time which extends into the distant future. Individual components of the model chains used can be tested to some extent, but how can the integrated assessment be evaluated to give additional confidence in the overall results? In order to address this issue, the Kristallin-l performance assessmentI has been compared with other HLW/spent fuel performance assessments. This technical comparison forms the basis of this paper which summarises the results from a more extensive comparison of a number of HLW/spent fuel disposal concepts, safety assessment methodologies and models2. It is part of a wider study which attempted to put the results of the Kristallin-I performance assessment into perspective by use of both this technical comparison with other performance assessments and by consideration of other information (e.g. natural analogue studies, environmental radiation data and risk assessments for radiological and non-radiological hazards) which allows the reasonableness of the results and significance of the predicted doses to be illustrated. Due to limitations of space, this aspect of the study has not been included here and the reader is referred to the “Kristallin-I: Results in Perspective” report2 for more details.

Radionuclides leaking from a canister spread into the backfill material surrounding the canister and then migrate through different pathways into the mobile water in the rock fractures. In order to quantify uncertainties regarding the release to the far field, a sensitive analysis is performed for the uncertainties in the parameters governing the release.

For nuclides escaping through a damaged canister in a repository of the type KBS-3 [1], most nuclides migrate through the pathways appointing to the fracture intersecting the deposition hole and the disturbed zone below the tunnel.

The sensitivity analysis showed that the grade of impact on the release of uncertainties in the parameters cannot be generalized for all nuclides, however, some patterns of the influence can be given. Uncertainties in the hole size and uranium solubility for nuclides embedded in the matrix have a strong influence on the release. The adsorption capacity of the bentonite surrounding the canister and the hydraulic properties of the fractured rock in the short pathways to the mobile water have to be accurately known to reduce uncertainties in the release to the far field. For nuclides with a high diffusion coefficient in the bentonite, e.g. cesium and strontium, the hydraulic properties of the fractured rock nearest to the canister have to be well known. For the iodide, on the other hand, with a very low diffusion coefficient, uncertainties in the hydraulic properties have a very small effect on the release. The uranium release with a very high inventory is not affected by uncertainties in the hydraulic properties of the rock.

The released nuclide fraction through each pathway is not largely influenced by uncertainties in the transport properties of the backfill materials whereas uncertainties in the water flow rates in the zones in the rock where the water flows, have a strong influence on the nuclide release distribution.

Performance/safety assessments have been conducted for the Waste Isolation Pilot Plant, a repository for transuranic waste in salt. We challenge two key assumptions in the 1992 Performance Assessment that led to non-conservative results. The predicted release rate of radionuclides appears underestimated due to the use of expert elicited solubilities. The overall release has been underestimated due to the use of improperly elicited human intrusion probabilities.

The repository concepts developed by PNC and Nagra for the disposal of vitrified high-level waste show many common features; both concepts involve deep geological disposal, with massive engineered barriers of similar design. PNC and Nagra have recently published comprehensive performance assessments based on their repository concepts and these are compared in order to gain assurance that the conceptual and mathematical models employed are state-of-the-art and to build confidence in the datasets in the assessment results. The predicted performance of the engineered barriers is more favourable to safety in the case of the Nagra Assessment, whereas the opposite is true of the geological barriers. These differences in predicted performance can be traced to a small number of differences in model assumptions and data. It is concluded that the differences are consistent with the somewhat different requirements of the two assessments and that the comparison enhances confidence in both assessments.

A very small number of disposal containers of heat generating nuclear waste may have initial manufacturing defects that would lead to pin-hole type failures at the time of or shortly after emplacement. For sufficiently long-lived containers, only the the initial defects need be considered in modelling of release rates from the disposal vault. Two approaches to modelling of near-field mass transport from a single point source within a disposal room have been compared: the finite-element code MOTIF (A Model Of Transport In Fractured/porous media) and a boundary integral method (BIM). These two approaches were found to give identical results for a simplified model of the disposal room without groundwater flow. MOTIF has been then used to study the effects of groundwater flow on the mass transport out of the emplacement room.

Sorptive properties of weathered dunites, gabbro-diabases and basic volcanic rocks for Sr and Cs were studied. The results show that the sorptive capacities of these rocks are equivalent to or, in some cases, superior to industrial sorptive materials. Results of a uranium distribution study by fission-track radiography suggest that material from weathered basic rocks is characterized by high sorptive properties for uranium also. One can assume that other radionuclides of the transuranic group will be intensely sor-bed by the residuum of weathered basic rocks. Low-temperature hyd-rothermal transformation leads to sealing fissures of the basic rocks with highly sorptive minerals, for example, smectite, chlorite, serpentine, talc, zeolite, hydroxides of Fe, Ti, Mn. The process results in contemporaneous decreasing hydraulic conductivity and increasing sorptive capacity of the rocks. HLW disposal at the radiochemical plant “Mayak” is expected to be produced in deep wells situated in basaltic rocks. The safety of disposal is based on high sorptive properties of the crust of weathering and protective capacities of volcanic rocks. This method is not expensive and may allow the disposal of HLW in the near future.

A new computer code, Analysis of RadionuclidESource-Term with Chemical Transport (AREST-CT), is described in this paper. The code is being designed to support performance assessment analyses of engineered systems for subsurface isolation of hazardous and radioactive wastes. Radionuclide releases from an engineered system are modeled by solving governing equations describing conservation of water mass, air mass, thermal energy, and chemical species mass. As such, the AREST-CT code will be capable of simulating radionuclide release and transport in a non-isothermal, unsaturated-saturated setting. Constituitive equations are implemented that describe corrosion of iron-based container materials, glass, and spent fuel waste forms. The governing equations are solved in a two-dimensional domain using an integrated finite-volume method. A third-order total variation diminishing (TVD) numerical scheme is evaluated to minimize numerical oscillations and dissipation of steep concentration gradients in advection-dominated transport problems.

UK Nirex Ltd is planning the deep geological disposal of intermediate- and low-level radioactive wastes. A site close to Sellafield in Cumbria in the north-west of England has been selected for evaluation, and an extensive programme of site characterization is underway. In support of this programme of characterization, and in preparation for presentation of a post-closure radiological safety case, performance assessment using mathematical modelling has been carried out by the Disposal Safety Assessment Team at AEA Technology, on behalf of UK Nirex Ltd. This paper describes recent developments of the assessment models relating to the groundwater pathway for return of radionuclides to the environment.

During the post-closure phase of a repository for low- and intermediate-level radioactive wastes significant quantities of gas will be generated, primarily by the anaerobic corrosion of metals and the degradation of organic wastes. A description is given of recent work carried out within the Nirex Safety Assessment Research Programme to address the generation of gas within a repository and the migration of gas from the repository to the biosphere. Theoretical modelling capabilities have been developed to address both of these issues. In order to gain confidence in such theoretical models, it is important to validate model predictions against observed experimental data. Preliminary experiments carried out to help validate the modelling approaches are described.

Dissolution of spent fuel has been studied in saline, anaerobe, carbonate free solutions. Processes controlling spent fuel dissolution and associated radionuclide release are radiolytically controlled oxidative dissolution, sorption on container, solubility and coprecipitation. Upper limits for oxidative dissolution rates are given by the production rates of oxidative radiolysis products. This limitation leads to a strong decrease in surface area normalized reaction rates with increasing surface to volume ratio (S/V) and imposes geometric constraints on prediction of spent fuel behavior in a repository. Solution concentrations of Am during spent fuel corrosion were about 5 orders of magnitude lower than the solubility of Am(OH)3(s) and are likely controlled by coprecipitation. Pu concentrations may be controlled by Pu(VI) or Pu(IV) (hydr)oxides.

The leaching behaviour of sintered UO2 pellets has been studied in synthetic groundwater at 96 °C under strictly controlled conditions (CO2 partial pressure, O2 partial pressure, pH and redox potential). Results obtained in oxidizing conditions (O2 content varying from 5 to 25 ppm) show a strong initial uranium release in solution during the first five days and a solubility control essentially driven by the growth of U(VI) species such as oxides, hydrates and carbonates. Results obtained in reducing conditions (use of a mineralogical buffer of Ni(0)/Ni(ll) with E0 = -0.257 Volts) show a very weak uranium release in solution close to 10-11 mole/l and a solubility control essentially driven by the formation of U(IV) compounds such as coffinite or uraninite.

Spent MTR fuel from German Material Test Reactors will not be reprocessed, but stored in a final salt repository in the deep geologic underground. Fuel elements will be placed in POLLUX containers, which are assumed to resist the corrosive attack of an accidentally formed concentrated salt brine for about 500 years. After a container failure the brine would contact the fuel element, corrode the aluminum plating and possibly leach radionuclides from the fuel.

A source term for the calculation of radionuclide mobilization results from the investigation of the behavior of MTR fuel in this scenario, which has to be considered for the long- term safety analysis of a deep mined rock salt repository.

Experiments with the different plating materials show that the considered aluminum alloys will not resist the corrosive attack of a brine solution, especially in the presence of iron, under the conditions in a drowned salt mine repository. Although differences in the corrosion rates of about two orders of magnitude were observed when applying different parameter sets, the deterioration must be considered to be almost instantaneous in geological terms.

Radionuclides are mobilized from irradiated MTR fuel, when the “meat” of the fuel element becomes accessible to the brine solution. It seems, however, that the radionuclides are effectively trapped by the aluminum hydroxide formed, as the activity concentrations in the brine solution soon reach a constant level with the progressing corrosion of the cladding aluminum. In the presence of iron a more significant initial release was observed, but also in this case an equilibrium activity seems to be reached as a consequence of radionuclide trapping.

We have carried out an experimental comparison study of the dissolution rates of unirradiated UO2 and SIMFUEL pellets and particles (100–300 μm) in a standard NaCI/NaHC03 solution, under oxidizing conditions. We have performed the experiments using batch and flow methodologies. Both methodologies gave similar results, indicating that the overall oxidation/dissolution process is the same in both cases. The results from the experiments indicate that under these conditions the dissolution process is both oxygen and bicarbonate promoted. The dissolution rates we obtained are: R=2.4 ± 0.8 mg U/m2 d for U02 and R= 0.17 ± 0.05 mg U/m2 d for SIMFUEL. The results of the experiments indicate that the dissolution rate under oxic conditions is clearly dependent on the number of U(VI) surface sites which for spent nuclear fuel is a function of the extent of radiolytic oxidation.

The dissolution of unirradiated UO2 (s) has been studied in NaCl and MgCl2 brines under both reducing and oxidizing conditions.

The initial uranium release under reducing conditions has been attributed to the dissolution of an initial oxidized layer. The final uranium concentrations have been modeled by using the PHRQPITZ computer program giving the solubility of the solid phase UO2 (s).

Under oxidizing conditions, the initial release is the sum of the oxidized layer dissolution and the oxidation/dissolution of the UO2. The release rates calculated are 1.4·10−5 mol d−1 m−2 in NaCl2-brine and 3.6·10−5 mol d−1 m−2 in MgCl2-brine. After the initial release, uranium concentration in the NaCI-brine reaches a constant value, which has been attributed to the formation of a secondary solid phase. In MgCl2-brine, the uranium concentration increases slowly indicating, in this case, no control by secondary phase formation.

Effects of alpha radiolysis of water on the corrosion of nuclear fuel (UO2) have been investigated in solutions at pH = 9.5, i.e., a value close to that expected in groundwaters at the depth of the disposal vault proposed in the Canadian nuclear fuel waste management program, CNFWMP. The corrosion potentials of UO2 electrodes exposed to the products of alpha radiolysis of water were monitored as a function of alpha flux and exposure time in a specially designed thin-layer cell. The oxidative dissolution rates of UO2 are calculated from the steady-state values of the corrosion potential using an electrochemical model. A procedure to predict the dissolution rate of used nuclear fuel in groundwater as a function of fuel cooling time is described, and illustrated by calculating the dissolution rates for the reference used fuel in the CNFWMP (Bruce CANDU reactor fuel, burnup 685 GJ/kg U). It is shown that the oxidative dissolution of used fuel in the CNFWMP will be important only for time periods ≤ 600 a at this burnup and assuming no decrease in pH.

Combined gap and grain-boundary inventories of 129I in 14 used CANDU fuel elements were measured by crushing and simultaneously leaching fuel segments for 4 h in a solution containing KI carrier. From analogy with previous work a near one-to-one correlation vas anticipated between the amount of stable Xe and the amount of 129I in the combined gap and grain-boundary regions of the fuel. However, the results showed that such a correlation was only apparent for low linear power rating (LLPR) fuels with an average linear power rating of 642 kV/m. For high linear power rating (HLPR) fuels (>44 kw/m), the 129I values were considerably smaller than expected. The combined gap and grain-boundary inventories of 129I in the 14 fuels tested varied from 1.8 to 11.01, with an average value of 3.6 ± 2.4% which suggests that the average value of 8.1 ± 1% used in safety assessment calculations overestimates the instant release fraction for 129I. Segments of used CANDU fuels were leached for 92 d (samples taken at 5, 28 and 92 d) to determine the kinetics of 129I release. Results could be fitted tentatively to half-order reaction kinetics, implying that 129I release is a diffusion-controlled process for LLPR fuels, and also for HLPR fuels, once the gap inventory has been leached. However, more data are needed over longer leaching periods to gain more understanding of the processes that control grain-boundary release of 129I from used CANDU fuel.

The long term behaviour of spent nuclear fuel is discussed in the light of recent thermodynamic and kinetic data on mineralogical analogues related to the key phases in the oxidative alteration of uraninite. The implications for the safety assessment of a repository of the established oxidative alteration sequence of the spent fuel matrix are illustrated with Pagoda calculations. The application to the kinetic and thermodynamic data to source term calculations indicates that the appearance and duration of the U(VI) oxyhydroxide transient is critical for the stability of the fuel matrix.

The leaching of the spent fuel matrix (UO2) is function of the radiolytic products formation. The effect of each radioiytic product on the leaching process is not totally understood. In the literature, the influence of H2O2 on the dissolution process is described from the qualitative point of view, and most of the studies were performed for pH values from 8 to 12. In this paper we report on the effect of the H2O2 in the leaching process of UO2 by dissolution experiments at various H2O2 concentrations. Also, it was tested the influence of S/V ratio (surface area exposed to the leaching media) on the UO2 leaching and secondary phases formation. It was identified the formation of secondary phases on the UO2 surface. Solid phases characterization was carried out by x-ray Photoelectron Spectrometry (XPS), x-ray Diffraction (XRD) and Scanning Electron Microscopy (SEM) techniques. By XPS studies the secondary phase formed corresponded to a U(VI) phase. By XRD analyses the solid was identified as studtite, UO4 - 4H2O, (card n0 16–206, [I]). A comparison of the U(VI) phases formed in spent fuel and UO, leaching experiments in various media has been carried out.