ATOMIC ENERGY &S& L'ENERGIE ATOMIQUE OF CANADA LIMITED \ A DU CANADA, LIMITEE

(1)

AECL-7796

ATOMIC ENERGY &S& L'ENERGIE ATOMIQUE

OF CANADA LIMITED

\

A

¥ DU CANADA, LIMITEE

EFFECT OF LOCALIZED WASTE-CONTAINER FAILURE ON RADIONUCLIDE

TRANSPORT FROM AN UNDERGROUND NUCLEAR WASTE VAULT

EFFETS EXERCES PAR UNE RUPTURE LOCALISEE DE CONTENEUR DE DECHETS SUR LE

TRANSPORT DES RADIONUCLEIDES A PARTIR D'UNE ENCEINTE SOUTERRAINE DE

DECHETS NUCLEATES

S. C. H. Cheung, T. Chan

Whiteshell Nuclear Research F*ch

;

ssement de recherches

Establishment nucleaires de Whiteshell

Pinawa, Manitoba ROE 1 LO

July 1983 juillet

(0

(2)

(3)

ATOMIC ENERGY OF CANADA LIMITED

EFFECT OF LOCALIZED WASTE-CONTAINER FAILURE ON RADIONUCLIDE

TRANSPORT FROM AN UNDERGROUND NUCLEAR WASTE VAULT

by

S.C.H. Cheung and T. Chan

Whiteshell Nuclear Research Establishment

Pinawa, Mav.itoba ROE 1L0

193••. J j l y

(4)

EFFETS EXERCÉS PAR UNE RUPTURE LOCALISÉE DE CONTENEUR DE DECHETS SUR LE TRANSPORT DES RADIONUCLÉIDES À PARTIR D'UNE ENCEINTE SOUTERRAINE DE

DÉCHETS NUCLÉAIRES par

S.C.H. Cheung et T. Chan

RÉSUMÉ

L'une des options de l'évacuation géologique des déchets de com-bustible nucléaire consiste à placer le conteneur de déchets dans un trou pratiqué dans le sol de l'enceinte souterraine. A l'intérieur de ce trou, on entoure le conteneur de terre compactée qu'on appelle tampon. On a pro-cédé â une simulation par éléments finis dans le but d'étudier les effets d'une rupture partielle localisée du conteneur de déchets sur le transport des radionucléides â l'état stationnaire par diffusion â partir du conteneur â travers le matériau vers la roche avoisinante et/ou les matériaux de rem-blayage. Dans la présente étude, on suppose que la concentration des radio-nucléides au niveau de l'interface "matériau tampon/matériau de remblayage" est nulle.

Deux cas sont analysés au niveau de l'interface entre le matériau tampon et la roche. Dans le premier cas, on a recours à une condition de limite sans flux pour simuler la roche intacte. Dans le deuxième cas, on utilise une concentration constante de radionucléides pour simuler la roche fissurée avec circulation d'eau souterraine. Selon les résultats obtenus, l'effet qu'exerce la rupture partielle localisée du conteneur de déchets sur le flux total dépend des conditions de limite â l'interface matériau tampon/ roche. Dans le cas de la roche intacte, le flux total dépend principalement de l'emplacement de la rupture. Plus l'emplacement de la rupture se situe vers le haut du conteneur mis en place, plus le flux total augmente. Dans le cas d'une rupture localisée donnée du conteneur de déchets, la surface de rupture se trouvant en dessous de la partie supérieure de la rupture n'a aucun effet sur le flux total. En ce qui concerne la roche fissurée, le flux total est directement proportionnel à la surface de rupture du conte-neur de déchets, quelque soit l'emplacement de la rupture.

L'Energie Atomique du Canada, Limitée

Établissement de recherches nucléaires de Whiteshell Pinawa, Manitoba ROE 1L0

1983 juillet

(5)

EFFECT OF LOCALIZED WASTE-CONTAINER FAILURE ON RADIONUCLIDE TRANSPORT FROM AN UNDERGROUND NUCLEAR WASTE VAULT

by

S.C.H. Cheung and T. Chan

ABSTRACT

In the geological disposal of nuclear fuel waste, one option is to emplace the waste container in a borehole drilled into the floor of the underground vault. In the borehole, the waste container is surrounded by a compacted soil material known as the buffer. A finite-element simulation has been performed to study the effect of localized partial failure of the waste container on the steady-state radionuclide transport by diffusion from the container through the buffer to the surrounding rock and/or backfill. In this study, the radionuclide concentration at the buffer-backfill interface is assumed to be zero.

Two cases are considered at the interface between the buffer and the rock. In case 1, a no-flux boundary condition is used to simulate intact rock. In case 2, a constant radionuclide concentration condition is used to simulate fractured rock with groundwater flow. The results show that the effect of localized partial failure of the waste container on the total flux is dependent on the boundary condition at the buffer-rock interface. For the intact rock condition, the total flux is mainly depend-ent on the location of the failure. The total flux increases as the loca-tion changes from the bottom to the top of the emplaced waste container. For a given localized failure of the waste container, the total flux re-mains unaffected by the area of failed surface below the top of the

failure. For fractured rock, the total flux is directly proportional to the failed surface area of the waste container regardless of the failure location.

Atomic Energy of Canada Limited Whiteshell Nuclear Research Establishment

Pinawa, Manitoba ROE 1L0 1983 July

(6)

1. INTRODUCTION

The Canadian Nuclear Fuel Waste Management Program is assessing^ the concept of underground disposal of radioactive fuel wastes from CANDU reactors in stable geological formations.

At present, disposal in a vault mined into a plutonic rock, such as a granite batholith of the Precambrian Shield, is receiving most

attention [1J. In such a vault the only potentially significant mechanism of radionuclide release is for circulating groundwater to penetrate to the waste, leach out the radionuclides, and carry them back to the surface. A number of protective barriers will be used to isolate the waste and mini-mize the probability of any significant escape. These are the waste form itself (either irradiated UC>2 fuel bundles, or solidified reprocessing waste if a CANDU fuel-recycle option is eventually adopted), the waste con-tainer, the buffer material surrounding the concon-tainer, the backfill and sealing materials which fill and seal the remainder of the vault, and the massive geological formation (see Figure 1 ) . The buffer and backfill are compacted clay/sand mixtures [2].

Two options for emplacement of the waste containers are currently being assessed [3]. These are in-room and borehole emplacement, as shown in Figure 2. In previous work, the near-field radionuclide transport for borehole emplacement was studied theoretically for the case where the waste container was assumed to have failed completely. At failure, all the con-tainer shell was assumed to be removed instantaneously from the system [4,5,6]. In the present paper, radionuclide transport is investigated for the case of localized partial failure of the waste container. This situa-tion is mor* representative of the expected mode of container failure.

2. NEAR-FIELD MASS TRANSPORT MODELS WITH LOCALIZED CONTAINER FAILURE

In these models, the following assumptions and boundary conditions are used:

2.1 ASSUMPTIONS

(1) Mass transport in the buffer is by diffusion only; i.e., there is no groundwater flow within the buffer. Since the buffer likely will have a very low hydraulic conductivity, this is expected to be a realistic approximation.

(2) The buffer is considered as a porous medium with a constant and isotropic ionic diffusion coefficient for all radionuclides. CANadian Deuterium Uranium (Canada's natural uranium-fuelled, heavy-water-moderated and cooled reactor).

(8)

2

-(3) There are no chemical reactions that affect radionuclide diffusion.

(4) Steady-state diffusion.

(5) The waste container is cylindrical and is placed symmetrically in the buffer material.

(6) The radionuclide concentration is constant at the failed area of the container surface.

2.2 BOUNDARY CONDITIONS

2.2.1 Boundary Conditions At Buffer-Backfill and Buffer-Rock Interfaces

The boundary conditions used in these models are shown in Figure 3. The configuration of the simulated system will remain unchanged throughout this study. At the buffer-backfill interface, the radionucllde concentration, C~, is assumed to be constant. The concentration, C^, de-pends mainly on the groundwater flow velocity in the backfill and the effective diffusion coefficients in the buffer and backfill [6]. For the buffer-rock interface, two cases are considered.

In case 1, shown in Figure 3(a), a no-flux condition is assumed to exist at the buffer-rock interface. This is a good approximation for the case of intact rock. In case 2, shown in Figure 3(b), a constant con-centration condition, C2, equal to that at the buffer-backfill interface is assumed to exist at the buffer-rock interface. This implies that the effective diffusion coefficient and groundwater flow velocity in the frac-tured rock are similar to those in the backfill.

2.2.2 Boundary Condition At Container-Buffer Interface

In the numerical model, a constant concentration, C j , is assumed to exist at the failed area of the container surface. This boundary con-dition is equivalent to assuming a constant source of radionuclides.

For each of the two cases discussed above, six different modes of container failure are considered, as illustrated in Figure 4. These

different modes of failure, labelled A to F in Figure 4, represent various possible locations and areas of the failed portion of the container sur-face . A and E indicate that a narrow annulus of the container sursur-face has failed. B and C represent failure of the bottom and top end surfaces of the container, respectively. D indicates that, half of the container, in-cluding the bottom face, has failed and F indicates complete failure. The unfailed area of the container is represented by a no-flux boundary

condition.

The total radionuclide flux through the buffer-backfill and the buffer-rock interfaces is evaluated for all the above cases.

(9)

3

-3. NUMERICAL METHOD

A finite-element program, DOT, [7] is used to simulate the dif-fusional transport of radionuclides• The DOT program solves the heat con-duction equation in two-dimensional axisymmetric (or planar) configurations with arbitrary boundary conditions. Since concentration diffusion and heat conduction are governed by the same type of partial differential equations, the DOT program is also applicable to the present problem.

The DOT program has been verified by comparison with analytical solutions for both steady-state and transient heat conduction under various boundary conditions [7]. It has also been validated against temperature data from a field experiment in which electrical heaters were emplaced in granitic rock to simulate the thermal loading from nuclear fuel waste in containers [8].

A. RESULTS AND DISCUSSIONS

Figure 5 shows the radionuclide concentration contours for the various container failure modes considered, for the intact rock condition. In view of the symmetry, only half longitudinal sections of a borehole emplacement configuration are presented. Each contour plot in Figure 5 corresponds to a particular failure identified in Figure 4. The quantity plotted is the normalized concentration difference, C = (C - C_)/(C. - C « ) , expressed as a percentage. Thus, at the failed portion at the container-buffer interface_C = Ct and C = 100%, while at the container-buffer-backfill inter-face C = C~ and C = 0%. For all cases in Figure 5, C below the top of the failed area is approximately 100% everywhere. The physical explanation is as follows. Initially, the radionuclide will diffuse both radially and vertically. Since a no-flux condition exists at the buffer-rock interface, the concentration in the buffer below the top of the failed area will in-crease, and the concentration gradient will dein-crease, with time. Eventu-ally the gradient becomes zero, and the radionuclide concentration is then uniform and equal to that at the failed area of the container surface. At this stage, steady-state is reached.

Above the failure location, the normalized concentration, C, de-creases from 100% at the failure location to 0% at the buffer-backfill interface.

Figure 6 shows the total normalized flux at the buffer-backfill interface as a function of the failed area (expressed as a percentage of the total container surface area) for the six cases illustrated in

Figure 5. The normalization of flux is obtained by using a unit diffusion coefficient and unit normalized concentration at the failed portion of the container surface. Although the failed areas are different for cases A and D, the total flux is found to be the same. Similarly, cases C, E and F, with different failed areas, also givi rise to the same total flux.

There-fore, it can be concluded that the total flux is independent of failure

(10)

4

-and E -and between B -and C, the total flux is seen to be greater the closer the failure occurs to the top of the container. This is because of the in-crease in concentration gradient when the failure occurs at a higher posi-tion on the container.

Figure 7 shows the percentage concentration contours of various modes of localized failure of the container for the fractured rock condi-tion. In general, the concentration decreases both radially and vertically from the failed portion of the container. Figure 8 shows the total normal-ized flux as a function of the failed area, (expressed as a percentage of the container surface area), for the six cases illustrated in Figure 7. This indicates that the total flux is almost linearly proportional to the failed surface area but independent of the location at which failure occurs.

Comparing Figures 6 and 8, the total flux is much lower for the intact rock condition than for the fractured rock condition, for the same failure geometry of the container.

5. ENGINEERING APPLICATIONS

Based on the above results, the design of the waste package system, consisting of the waste container and buffer, should take the rock properties into account. In the case of intact rock, the buffer material above the waste container should be engineered to have a very low diffusion coefficient and to have as large a thickness as practical. Furthermore, the top end surface of the container should, in principle, be designed to be the most corrosion resistant part of the container. In the case of fractured rock, the radial buffer thickness, i.e., the thickness of the annulus between the container and the rock, should be maximized and the diffusion coefficient of the buffer should be minimized. However, it has been shown [6] that increasing the radial buffer thickness beyond 0.45 m and the vertical buffer thickness above the waste container beyond 1 m would have very little additional effect in reducing the total flux. Since the total flux is proportional to the failed area, the container should also be designed to have as small a radial surface area as is practical for a given volume of waste.

6. CONCLUSIONS

The following conclusions can be drawn from the above studies: (1) For the intact rock condition, the total flux depends primarily

on the location of the failure on the waste container, but is practically independent of the failed surface area.

(2) For fractured rock, the total flux varies directly with the failed surface area, regardless of its location.

(11)

5 -ACKNOWLEDGMENTS

Reviews by V. Guvanasen, R.S. Lopez and D.B. McConnell substan-tially improved this manuscript.

REFERENCES

1. J. Boulton, (Editor) "Management of Radioactive Fuel Wastes: The Canadian Disposal Program," Atomic Energy of Canada Limited Report, AECL-6314 (1978).

2. G.W. Bird and D.J. Cameron, "Vault Sealing Research for the Canadian Nuclear Fuel Waste Management Program," Atomic Energy of Canada Limited Technical Record, TR-145* (1982).

3. S.C.H. Cheung, " A Comparison of Borehole and In-room Emplacement for. a Canadian Nuclear Waste Disposal Vault," Atomic Energy of Canada Limited Technical Record* (in preparation).

4. S. Page and S.C.H. Cheung, "Diffusional Mass Transport Phenomena in the Buffer Material and Damaged Zone of a Borehole Wall in an Underground Nuclear Waste Vault," Atomic Energy of Canada Limited Report, AECL-7788 (1983).

5. S.C.H. Cheung and T. Chan, "Parameter-Sensitivity Analysis of Near-Field Radionuclide Transport," Atomic Energy of Canada Limited Report (in preparation).

6. S.C.H. Cheung, G.W. Bird and C.B. So, "Diffusion Modelling of the Borehole Emplacement Concept for a Nuclear Fuel Waste Disposal Vault," Proceedings of the NEA Workshop on Near Field Phenomena in Geological Repositories for Radioactive Waste. NEA/OED, Paris (1981), pp. 375-388.

7. R.M. Polivka and E.L. Wilson, "Finite Element Analysis of Nonlinear Heat Transfer Problems," Report No. UCSESM 76-2, Department of Civil Engineering, University of California, Berkeley, (1976).

8. T. Chan, I. Javandel and P.A. Witherspoon, "Heat Transfer in Underground Heating Experiments in Granite, Stripa, Sweden," Proceedings of the Technical Session on Heat Transfer in Nuclear Waste Disposal, ASME 1980 Winter Annual Meeting, Chicago, np 1-8, also issued ab Lawrence Berkeley Laboratory Report LBL-10876 (1980).

* Unrestricted, unpublished report available from SDDO, Atomic Energy of Canada Limited Research Company, Chalk River, Ontario KOJ 1J0.

(12)

6

-Mm MASSIVE GEOLOGIC

Wi& FORMATION

BACKFILLED

EXCAVATION

BUFFER MATERIAL

OURABLE

CONTAINER

SOLIDIFIED

NUCLEAR

WASTE

FIGURE 1: Schematic Illustration of the Sequence of Barriers to the Release of Radionuclides in a Nuclear Fuel Waste Disposal Vault

(13)

7

-BACKFILL

(a)

IN-ROOM

EMPLACEMENT

(b)

BOREHOLE

EMPLACEMENT

BACKFILL

BUFFER

WASTE

CONTAINER

FIGURE 2: Schematic Illustration of Borehole and In-Room Emplacement Concepts (not to scale)

(14)

BACKFILL

Ca

BACKFILL

C

2

INTACT^

ROCK ^

^

/

///

BUFFER

\

a:

LU

?

o

\ ,

/

/INTACT FRACTURED

^ROCK ROCK

/

/ C

2

/

BUFFER

a:

UJ

o

c,

FRACTURED

ROCK

C

2

i 03

NO FLUX

CASE (I )

(a)

C

2

CASE (2)

(b)

FIGURE 3: Schematic Illustration of Two Cases of Boundary Conditions at Buffer-Container, Buffer-Rock and Buffer-Backfill Interfaces

(15)

LEGEND

(A)

(B)

(C)

(E)

0000

FAILURE

SURFACE AREA

STRIP

PARTIAL SURFACE

END SURFACE

FIGURE 4: Schematic Illustration of Various Locations and Failed Areas of the Cylindrical Waste Container Surface

(16)

10

-100:

20|- 5 j

I § — i

I

r

FIGURE 5: Normalized Concentration Contours for Six Different Failure Modes of the Container for Intact Rock Condition

(17)

11

-X

0.7

0.6

0.5

ID

0.4

0.3

OC 0.2

o

0.1

CASKS

A

B

C

D

E:

F

SYMBOL

•

o

A

a

20 40 60 80 100

FAILED AREA OF THE CONTAINER SURFACE (%)

FIGURE 6: Total Normalized Flux as a Function of the Failed Area

Percentage of the Container Surface for Six Different Localized Failure Modes of the Container for Intact Rock

(18)

12 -.Or 0% 00% *#rov 01874 100* — 9 9 — , I 0187} lai J... oian (01 —100-I OlBTS IFI

FIGURE 7: Normalized Concentration Contours for Six Different Localized Failure Modes of the Container for Fractured Rock

(19)

13

-X

i

Q

UJ

N

35

30

25

20

-15 •

O 10

-1 i -1 • / / 1 1 1 1

a

CASES SYMBOL

A •

B *

C

-D o

E

A i

-20 40 60 80 100

FAILED AREA OF THE CONTAINER SURFACE(%)

FIGURE 8: Total Normalized Flux as a Function of the Failed Area

Percentage of the Container Surface for Six Different Localized Failure Modes of the Container for Fractured Rock

(20)

ISSN 0067-0367 ISSN 0067-0367

To identify individual documents in the series Pour identifier les rapports individuels faisani panic de cette we have assigned an AECL- number to each. serie nous avons assigne un numero AECL- "a chacun.

Please refer to the AECL- number when Veuillez faire mention du numero AECL- si vous requesting additional copies of this document demandez d'autres exemplaires de ce rapport

from au

Scientific Document Distribution Office Service de Distribution des Documents Officiels Atomic Energy of Canada Limited L'Energie Atomique du Canada Limitee

Chalk River, Ontario, Canada Chalk River, Ontario, Canada