A compacted mixture of Kunigeru VI bentonite and D-sand is being consideredfor use as an engineered barrier in low-level radioactive waste disposalfacilities in Japan. An important issue is the maintenance of theretardation property of the mixture during shear that might be induced insuch barriers by earthquakes and/or gradual tectonic deformations occurringover the design life of the facility. To investigate this issue, comparativetests on a bentonite-sand mixture and kaolin-sand mixture were conducted bymeans of a recently-developed coupled shear and permeability testingapparatus under temperature controlled condition. In addition topermeability, the specific storage of bentonite-sand specimen during shearis also systematically evaluated with the new analytical theory for theconstant flow permeability test. The present study reveals that 1)temperature control is preferred for measuring the permeability ofextremely-low permeability materials with the constant-flow pump method; 2)both the permeability and specific storage of the mixture of Kunigeru VIbentonite and D-sand (bentonite to sand weight ratio = 15: 85) were notsignificantly influenced by shear strains up to 3% whereas the permeabilityof the kaolin-sand mixture increased almost linearly with the increment ofshear strain; 3) the swelling of bentonite in the mixture under lowconfining stress decreases both the permeability and specific storage ofbentonite-sand mixture. This feature of bentonite helps maintain thelong-term retardation property of the mixture; and 4) the constant flowpermeability test method, with the newly derived theoretical analysis,promises to become a very effective means of investigating, rapidly andsystematically, the permeability and specific storage of extremely-lowpermeability materials with relatively-low (i.e. close to in situ) hydraulicgradients. This is necessary for the design and long-term performanceassessment of engineered barriers.

Compacted clay-based buffer surrounds corrosion-resistant waste containersin the Canadian concept for nuclear fuel waste disposal. Clays naturallycontain small quantities of organic matter that may be resistant tobacterial degradation. The containers with highly radioactive material wouldsubject the surrounding buffer to both heat (max. 95°C) and radiation (max.52 Gy/h). Both could potentially break down complex organic material tosmaller, more bioavailable compounds. This could stimulate microbial growthand possibly affect gas production, microbially-influenced corrosion orradionuclide migration. Experiments were carried out in which buffer (50wt.% Na-bentonite, 50 wt.% sand) was heated at 60 and 90°C for periods of 2,4 and 6 weeks, in some cases followed by irradiation to 25 kGy. Unheatedbuffer was also irradiated to 25 and 50 kGy at different moisture contents.The treated materials were subsequently suspended in distilled water, shakenfor 24 h and centrifuged to remove the solids. The 0.22 μm filter-sterilizedleachates were inoculated with equal volumes of fresh groundwater andincubated at room temperature for 10 d to determine the increase in totaland viable bacteria compared to a groundwater control. Results indicatedthat leachates from buffer subjected to heat, radiation or combinations ofthese, had a stimulating effect on both total and viable cell counts ingroundwater, compared to unamended groundwater controls. This stimulatingeffect was generally most pronounced for viable counts and could be largerthan two orders of magnitude. Leachates from untreated buffer material alsostimulated the growth of groundwater bacteria, but to a lesser extent thanleachates from heat-and radiation-treated buffer material. The effects ofheat and radiation on nutrient availability in clay-based sealing materials(at relevant clay/moisture ratios) should, therefore, be taken into accountwhen attempting to quantify the effects of microbial activity on vaultperformance.

In this study, gas migration experiments in unsaturated and saturated stateswere carried out to clarify the fundamental gas migration characteristics incompacted bentonite to be used for the geological disposal of high-levelradioactive waste. In unsaturated experiments, the gas permeability forJapanese bentonite (Kunigel VI) as a function of degree of saturation wasmeasured to examine the applicability of conventional two-phase flow modelsto compacted bentonite. The intrinsic permeability obtained in this studywas about five orders of magnitude larger than that obtained in waterpermeation tests with the same density. The difference seems to originatefrom the change of pore structure due to the swelling phenomenon of thebentonite. Since these effects have not been evaluated quantitatively yet,various relative gas permeability functions of conventional two-phase flowmodels were applied as a first approximation.

Saturated experiments designed to simulate the gas migration phenomenon in arepository for the waste were carried out to obtain relationship betweenbreakthrough and swelling pressures using Kunigel VI and French Fo-Ca clayin saturation state. The reproducibility of the breakthrough pressure wasalso examined for Kunigel VI bentonite. The breakthrough pressure was almostthe same as swelling pressure irrespective of the type of clay. As to thereproducibility of breakthrough pressure, it was observed that first andsecond breakthrough pressures were almost the same for Kunigel VI specimenswith the dry densities of 1.7 and 1.8 g/cm3.

Controlled flow-rate gas injection experiments have been performed onpre-compacted samples of KBS-3 specification M×801 bufferbentonite using helium as a safe replacement for hydrogen. By simultaneouslyapplying a confining pressure and backpressure, specimens wereisotropically-consolidated and fully water-saturated under pre-determinedeffective stress conditions, before injecting gas using a syringe pump.Ingoing and outgoing gas fluxes were monitored. All tests exhibited aconspicuous threshold pressure for breakthrough, somewhat larger than thesum of the swelling pressure and the backpressure. All tests showed apost-peak negative transient leading to steady-state gas flow. Using astepped history of flow rate, the flow law was shown to be nonlinear. Withthe injection pump stationary (i.e. zero applied flow rate), gas pressuredeclined with time to a finite value. When gas flow was reestablished, thethreshold value for gas breakthrough was found to be significantly lowerthan in virgin clay. There is strong evidence to suggest that the capillarypressure for the penetration of interparticle pore space of buffer bentoniteis of such a magnitude that normal two-phase flow is impossible. Gas entryand breakthrough is therefore accompanied by the development of microcrackswhich propagate through the clay from gas source to sink. The experimentssuggest that these pathways open under high gas pressure conditions andpartially close if gas pressure falls, providing a possible explanation ofthe nonlinearity of the flow law.

Laboratory experiments were performed to study the interaction betweengroundwater and compacted sodium bentonite (Volclay MX-80). The solutionsused were the fresh and saline groundwater simulants. The experiments werecarried out in aerobic and anaerobic conditions at elevated temperature. Ofmain interest in the present study were the chemical changes in the reactingsolution, bentonite porewater. and bentonite itself. The results for majorcations display a principal difference between the interactions with freshand saline solutions, while the differences between aerobic and anaerobicconditions within each solution case seem to be minor. The experimentalresults for the bentonite-water equilibria were interpreted in terms of amulti-site surface complexation model and the computer program HYDRAQL. Theapparent diffusivities for sodium and sulphate in bentonite samplessandwiched between two filterplates were also determined.

Gas migration properties of bentonite/sand mixtures for a large-scalestructure were estimated on the basis of laboratory test results obtainedwith small test specimens. Water and gas migration tests were carried out toclarify scale effects on migration properties of the mixtures.

Three-dimensional calculations that explicitly represent a realistic mixtureof waste packages (WPs) are used to analyze decay-heat-driventhermal-hydrological behavior around emplacement drifts in a potentialhigh-level waste facility at Yucca Mountain. Calculations, using the NUFTcode, compare two fundamentally different ways that WPs can be arranged inthe repository, with a focus on temperature, relative humidity, andliquid-phase flux on WPs. These quantities strongly affect WP integrity andthe mobilization and release of radionuclides from WPs. Point-load spacing,which places the WPs roughly equidistant from each other, thermally isolatesWPs from each other, causing large variability in temperature, relativehumidity, and liquid-phase flux along the drifts. Line-load spacing, whichplaces WPs nearly end to end in widely spaced drifts, results in morelocally intensive and uniform heating along the drifts, causing hotter,drier, and more uniform conditions. A larger and more persistent reductionin relative humidity on WPs occurs if the drifts are backfilled with alow-thermal-conductivity granular material with hydrologie properties thatminimize moisture wicking.

A coupled model concept which may be used for performance assessment of anuclear repository is presented. The tool is developed by integration of twomodels, one near field and one far field model. A compartment model,NUCTRAN, is used to calculate the near field release from a damagedcanister. The far field transport through fractured rock is simulated byusing CHAN3D, based on a three-dimensional stochastic channel networkconcept. The near field release depends on the local hydraulic properties ofthe far field. The transport in the far field in turn depends on where thedamaged canister(s) is located. The very large heterogeneities in the rockmass makes it necessary to study both the near field release properties andthe location of release at the same time. In order to demonstrate thecapabilities of the coupled model concept it is applied on a hypotheticalrepository located at the Hard Rock Laboratory in Äspö, Sweden. Two mainitems were studied; the location of a damaged canister in relation tofracture zones and the barrier function of the host rock. In the study ofthe near field rock as a transport barrier the effect of different tunnelexcavation methods which may influence the damage level of the rock aroundthe tunnel was addressed.

In a safety analysis of a repository for high level nuclear waste, it is ofprimary importance to identify and document all processes that are acting onthe repository system. An interaction matrix methodology have been appliedto the spent fuel subsystem. The purpose of this application is to identify,structure and rank the Process System and to discuss how the identifiedprocesses can be treated.

A general, integrated performance assessment code, AREST-CT, was used toanalyze the influence of various factors on the release rates ofradionuclides from a proposed facility for disposal of low-activity tankwastes. The code couples various process models together based on theframework of reaction-transport theory. The disposal facility was modeled asa 1-D column surrounded by soil. A borosilicate waste glass, LD6–5412 wasthe waste form considered in the analysis. Included in the simulations were38 aqueous species, 14 minerals, 21 equilibrium reactions, and 16 kineticreactions. Dissolution rate of the glass and the release rates of Te, Pu, U,Np, I, Se under different conditions were calculated for 50,000 years. Thesimulations revealed that 1) open exchange between the atmosphere andpore-water within the vault significantly improves the performance; 2) anion-exchange reaction between the glass and aqueous phase increases therelease rates significantly; and 3) at the hydrogeologie conditions underconsideration, variation of the pore-water velocity has little effect on therelease rate of radionuclides. These results provide a scientific basis forformulation of waste forms and engineering design of the disposal facility.Reaction-transport modeling can provide information on the long-termperformance of disposal systems that are not obtainable from laboratoryexperiments alone or by conventional decoupled process models.

A new mechanistic low-level waste source term model was developed. Keyfeatures of this effort are: use of cumulative probability functions todescribe the failures of waste containers; capability to describe diffusioncontrolled release which is dependent on the conditions of the waste packagesurroundings; consideration of the effects of gas generation on the sourceterm, and; use of a source inventory characterization routine to provide abuilt-in capability for defining the distributions of radionuclides invarious waste forms and streams. The model is capable of describing thediffusion of radionuclides in waste forms with the use of concentration andflux continuity boundary conditions. Release of radionuclides from variouswaste streams is modeled by the combination of diffusion, dissolution, andsurface release. Failures of waste containers are portrayed by the use ofprobabilistic failure functions based on Weibull and lognormaldistributions. The model for radionuclide release from waste packages iscoupled with the near-field transport model to describe the effects ofmigration and dispersion within the disposal unit. Characteristics of thenew model were evaluated through sensitivity analysis and compared withexisting source term codes.

A defense-in-depth engineered barrier system (EBS) is employed in thecurrent design concept for the potential high-level nuclear waste repositoryat Yucca Mountain, Nevada, USA. Simplifying the geometry of the cylindricalwaste container into the equivalent spherical configuration, andincorporating detailed analysis of the mechanics of water flow around thewaste container surface, a mathematical model is developed for advectiverelease from a “failed” (or perforated) waste container under drippingwater. It is shown that the advective release rates are controlled bydiffusion through the perforations in the waste container, and affected insignificantly by the dripping flow rate for the flowrate range considered. The release rates depend strongly on the number ofperforations (or pit penetrations) in the waste container. The insensitivityof the release rate to the dripping flow rate is explained by the fact thatradionuclide is released from the container surface to the boundary layer ofthe water film which contacts the container surface and is relativelystagnant. Also, since a laminar flow around the waste container surface isassumed in the model development, radionuclides transport across the waterlayers in the film by diffusion only. Additionally, the insignificant effectof the flow rate is contributed by the “short” penetration depth ofradionuclide into the water film assumed in the model development. Theanalyses show that the number of perforations, the size of perforation, thecontainer wall thickness, and the geometry (i.e., radius) of the wastecontainer are important parameters that control the advective release rate.It is emphasized that the “failed” (or perforated) waste package containercan still perform as a potentially important barrier to radionucliderelease.

A two-layer waste package (carbon steel outer barrier and Alloy 825 innerbarrier) is specified to dispose of high-level nuclear waste at thepotential repository at Yucca Mountain. A set of improvements and morerealism have been added to a stochastic waste-package degradation modelwhich was developed for a recent total system performance assessment of thepotential repository [1]. The waste-package surface is divided into“patches” to better represent the general corrosion of the carbon-steelouter barrier. The “corrosion-time” concept is developed to represent thecorrosion of the carbon-steel outer barrier in changing exposure conditionswith time such as those expected in the potential repository. With thepatches approach and the corrosion-time concept implemented into thewaste-package degradation model, sensitivity of the waste packagedegradation (failure and pitting degradation) to different thresholdspalling thicknesses of the corrosion products from the carbon-steel outerbarrier is analyzed. The results show that the waste-package pittingdegradation is sensitive to the corrosion-products spalling thickness of thecarbon-steel outer barrier. A greater pitting degradation of the wastepackages is predicted with a smaller spalling thickness. Furtherunderstanding of the corrosion-products spalling in different exposureconditions (i.e., water chemistry, water contact mode, etc.) and its effectson carbon steel corrosion is needed to enhance the confidence in thewaste-package performance modeling in the potential repository.

The Phase 3 Total System Performance Assessment (TSPA) sponsored by theElectric Power Research Institute (EPRI) has led to new insights intocritical models and parameters affecting estimated doses to humans from apotential repository of high-level radioactive wastes at Yucca Mountain,Nevada. The Phase 3 model has been extended to encompass time-varyingclimate and infiltration, detailed modeling of the source term andhydrology, and detailed specification of possible interaction betweenpercolating ground water and waste containers. The model estimates doses toa time of one million years.

The three key radionuclides contributing to estimated total doses areTc-99,1–129, and Np-237. Five other nuclides contributing to dose in lesser(but significant) amounts are U-233, Th-229, Pa-231, Ac-227, and Se-79.These results are consistent with other TSPAs.

From sensitivity studies, the most critical models and parameters are asfollows. Infiltration and percolation assumptions, including the amount oflateral diversion of infiltration water, are important and need verificationwith site data and/or more detailed modeling. Parameters of the unsaturatedzone (UZ) and saturated zone (SZ) determine dilution and delay ofconcentrations and peak doses downstream. The fraction of containers thatbecome wet are not critical in our model, but this lack of sensitivityreflects our coupling of the fraction with a model of focused flow past thecontainers; an different model might indicate higher sensitivity. Also, thedegree of coupling between fracture and matrix flow is important inaffecting the times of peak doses but not their magnitudes.

Other critical design assumptions that could lead to reduced and/or delayeddoses are a more robust container design, a capillary barrier around eachcontainer, the dilution during hydrologie transport from the repository to apotential agricultural community downstream, and the characteristics of an“average” individual in that community who might receive a dose.

In performance assessment of geological disposal systems for High LevelRadioactive Waste (HLW), the change of environment over the long-term mustbe considered. Therefore, it is necessary to consider a wide range ofparameters concerned with radionuclides migration, especially the dependenceof solubility on geochemical environment. In this study, assuming that therelease rate of the nuclides from buffer material is limited by inventoryultimately, the relationship between the initial inventory and thesolubility that produces a solubility-invariant maximum release rate fromthe buffer is examined by using a simple steady-state analytical solutionwithout decay. As the result, the threshold of “effective” solubility in theperformance assessment of the geological disposal systems for HLW isobtained as a function of initial inventory, distribution coefficient (Kd),diffusion coefficient, and thickness, porosity and density of the buffer.Also, the threshold of “effective” steady dissolution rate corresponding tothe threshold of “effective” solubility is obtained.

Recent proposals for a new U.S. standard for high-level waste disposal wouldlimit the average dose to individuals in the vicinity surrounding a geologicrepository. This would be a new approach to protecting the public fromenvironmental releases of radioactivity. Heretofore, criteria adopted forgeologic disposal have limited the reasonable maximum exposure to a futurehypothetical individual. Here we present quantitative analyses of therelation between maximum exposure and vicinity-average exposure, resultingfrom future human use of ground water contaminated by radioactive releasesfrom a repository.

Estimating the vicinity-average exposure would require postulates andguesses of location and habits of future people. Exposure probabilitiespostulated by others show that proposed dose limit to the vicinity-averageindividual would be a far more lenient standard than the traditional doselimit to reasonably maximally exposed individuals. The proposedvicinity-average dose limit would allow far greater concentrations ofcontaminants in ground water than would be allowed by normal standards ofground water protection. A safety standard that limits vicinity-averageexposure should also include limits on maximum exposure.

Thermal/hydrological modeling of the Radioactive Scrap and Waste Facility(RSWF) has been completed with the TOUGH2 (Transport of UnsaturatedGroundwater and Heat) Code.[1] The RSWF will be utilized as an interimstorage facility for ceramic and metallic waste forms developed from theelectrometallurgical treatment of spent nuclear fuel. The RSWF is an arrayof 1,350 carbon steel liners located at grade near Argonne NationalLaboratory-West on the Idaho National Engineering Laboratory (INEL).

The primary driving force for this modeling research was to assess thermalcapacity limits for RSWF liners so that heat generating materials can besafely stored. Maximum wasteform temperatures will be governed by both theamount of heat the system can dissipate and the orientation andcharacteristics of the wasteform. The focus of this report is on the amountof heat the interim storage system can safely dissipate. The effect of thetemporal variation of soil moisture on the performance of the RSWF isassessed. The facility was analyzed to determine the maximum allowablethermal loading of the RSWF liners.

A relatively simple model is suggested for radionuclide dispersion andassessment of additional population dose and radiation risk for radionucliderelease from the nuclear submanne (NS) #601 with Pb-Bi coolant that wasdumped in the Kara Sea.

Bias is a difference between model and reality. Bias can be introduced atany stage of the modelling process during a site characterisation orperformance assessment programme. It is desirable to understand such bias soas to be able to optimally design and interpret a site characterisationprogramme. The objective of this study was to examine the source and effectof bias due to the assumptions modellers have to make because reality cannotbe fully characterised in the prediction of ground-water fluxes. Awell-defined synthetic “reality” was therefore constructed for this study. Alimited subset of these data were independently interpreted and used tocompute groundwater fluxes across specified boundaries in a cross section.The modelling results were compared to the “true” solutions derived usingthe full dataset. This study clarified and identified the large number ofassumptions and judgements which have to be made when modelling a limitedsite characterisation dataset. It is concluded that bias is introduced ateach modelling stage, and that it is not necessarily detectable by themodellers even if multiple runs with varied parameter values areundertaken.

In parallel to studies of disposal of vitrified high-level waste fromreprocessing, projects have been initiated to examine options for directdisposal of spent fuel in Switzerland. The basic concept involves in-tunnelemplacement of encapsulated spent fuel in a deep repository which isbackfilled with compacted bentonite. Two possible host rocks are considered- crystalline basement and Opalinus Clay. This paper reports the results ofa thermal analysis which was carried out to evaluate constraints onrepository layout set by the desire to limit temperatures experienced by thebentonite backfill.