Bergaya F. Handbook of Clay Science

Подождите немного. Документ загружается.

11.4.1. BUFFER MATERIALS

Table 11.4.1 shows the candidate buffer-clay materials tested in international

research work, and the major physical properties that have been determined. Typical

hydraulic conductivity and swellin g pressure data of tested buffer materials are

collected in Tables 11.4.2 and 11.4.3 for different bulk densities.

In a few cases, the raw bentonite is so uniform and rich in smectite that only

simple drying in the sun of the excavated material with subsequent grinding is

required at the quarries and plants. However, in most cases the clay material must be

dried in rotating kilns, and also transform ed to a suitable Na

-form.

The processed clay is commonly stored in silos or in stock piles protected from rain,

and is packed in 25–50 kg paper bags or 1000 kg big bags for delivery. The preparation

of buffer clay is made by compaction of dry clay powder to blocks with a high density.

The blocks are then placed in the deposition holes to surround the canisters as

illustrated in Fig. 11.4.3 for the Swedish KBS-3 concept. Fig. 11.4.4 shows a big ring-

shaped block of bentonite for use in a KBS-3 repository (Johannesson et al., 1995).

Fig. 11.4.1. Repository concepts, most of them implying that canisters with highly radioac-

tive waste are surrounded by dense smectite clay and placed in holes or tunnels.

Chapter 11.4: Clays and Nuclear Waste Management704

For isolation of less dangerous radioactive waste with a lifetime of less than about 500

years and causing almost no heat by radioactive decay, loosely ﬁlled granulates of

smectite clay can be used as shown in Fig. 11.4.2.

11.4.2. PREDICTED PERFORMANCE OF THE ‘‘CLAY BUFFER’’

There are three major phases in the performance of the engineered buffer barrier:

a ﬁrst phase comprising application and evolution of the buffer including water

uptake and swelling, leading to maturation (0–100 years); a second phase with the

matured buffer exposed to a thermal gradient (100–1000 years); and a third phase

with tectonic or glacial impact under ambient or reduced temperature

(1000–1,000,000 years).

The ﬁrst phase involves strong radiation and complex physico-chemical processes

partly controlling the long-term performance of the buffer. This phase was regarded

30 m

50 m

Fig. 11.4.2. Schematic section through a big concrete silo for ﬁnal storage of low and in-

termediate radioactive waste in Sweden. (A) Cement-stabilised sand, (B) Concrete plugs, (C)

Top cover of bentonite/sand mixture, (D) Concrete silo, (E) Fill of granulated bentonite, (F)

Waste packages, (G) Bottom bed of bentonite/sand mixture (10/90) and (H) Rock drainage

with discharge to a lower tunnels system until the repository is closed.

11.4.2. Predicted Performance of the ‘‘Clay Buffer’’ 705

as particularly important and has therefore been the focus of most of the interna-

tional work up to now. If the hydraulic conductivity and swelling pressure are not

significantly changed during this phase, the buffer will be practically impermeable. As

such, it can serve as an excellent isolation of the waste in any time perspective. This

applies if the canisters stay intact for the ﬁrst few thousands of years, and significant

mineral changes in the buffer do not occur in any of the phases. It is therefore

of fundamental importance that the canister-embedding clay offers physical and

Table 11.4.1. Physical properties of some buffer clay candidates. The ﬁrst column indicates

the major smectite mineral (Mt. means Montmorillonite) and the name of the commercial

product

Clay Manufacturing company Smectite

content

Properties

KDp

l (f, a, b)

Mt. MX-80 American Colloid (USA) 75 * ****

Mt. Tixoton Su

d-Chemie (Germany) 90 * — * — —

Mt. Moosburg Su

d-Chemie (Germany) 65 * — * — *

Mt. IBECO Na Silver & Baryte (Greece) 80 * — * * *

Mt. IBECO Ca Silver & Baryte (Greece) 80 * — * * *

Saponite (Geohellas, Greece) 80 * — * — —

Beidellite (Through Enresa, Spain) 35 * — * — —

Kunigel Kunimine Ind., Japan 50 * — * — —

Montm. RMN Obrnice (Czech Republic) 90 * — * * —

Mixed-layer,

Friedland

DURTEC (Germany) 45 * — * — —

* ¼ tests performed

K ¼ Hydraulic conductivity, D ¼ Diffusivity, p

¼ Swelling pressure, l ¼ heat conductivity and capacity,

(f,a,b) ¼ rheological parameters.

Table 11.4.2. Hydraulic conductivity in m/s of SKB-tested potential buffer materials satu-

rated and percolated with distilled water. Bulk densities 1800–2100 kg/m

Density (kg/m

) 1800 1900 2000 2100

MX-80 4 10

13

210

13

810

14

310

14

IBECO, Na 9 10

13

—410

13

—

IBECO, Ca 2 10

11

—210

13

—

RMN — 4 10

13

—810

12a

Kunigel 3 10

12

—10

12

—

Beidellite 5 10

12

—510

13

—

Saponite 10

12

—510

13

—

With 5% graphite and 10% quartz powder.

Chapter 11.4: Clays and Nuclear Waste Management706

Fig. 11.4.3. Cross section of a deposition tunnel and hole for disposal of 20 ton canister of

copper and steel containing used reactor fuel rods. KBS-3 concept.

Table 11.4.3. Swelling pressure (p

) in MPa of a number of tested buffer materials saturated

with distilled water

Density at saturation (kg/m

) 1300 1500 1700 1800 1900 2000 2100

MX-80 0.06 0.2 0.4 0.8–0.9 1.4 4–5 10–12

IBECO, Na — — — 0.6–1 — 4–5 —

IBECO, Ca — — — 0.2 — 5 —

RMN — — 0.45 1.9

2.5

3.9

—

Beidellite — — — 1.5 — 4.2 —

Saponite — — — 2.5 — 8.8 —

Kunigel — — — 0.2 — 0.9 —

Friedland — — 0.05 0.1 0.3 0.8–1 2–2.5

Density at saturation 1850 kg/m

Mixture with 5% graphite and 10% quartz powder. Density at saturation 1950 kg/m

11.4.2. Predicted Performance of the ‘‘Clay Buffer’’ 707

chemical environments that preserve the tightness of the canisters for a long time, and

delay the migration of radionuclides that may ultimately escape from them.

Taking the KBS-3 reporitory concept as a reference for isolation of highly

radioactive waste by the use of compacted smectitic blocks a generalised conceptual

model of the buffer has the following features:

1. Thermally induced redistribution of the initial pore-water in the clay

2. Uptake of water from the rock and backﬁll leading to hydration of the buffer

3. Expansion of the buffer yielding displacement of the canisters and overlying

backﬁll

4. Dissolution of buffer minerals and precipitation of chemical compounds in the

buffer.

Processes related to temperature are termed (T); those related to hydraulic proc-

esses (H); those dealing with mechanical, or rather rheological, behaviour (M); and

mechanisms that have to do with chemical changes are termed (C).

A. Thermally Induced Redistribution of the Initial Pore-Water and Maturation of the

Pellet Fill(TH)

Pore-water is transported in liquid and vapour form from the hotter to the colder

parts of the buffer. This yields early saturation of the buffer close to the rock and

desiccation close to the hot canister.

Fig. 11.4.4. Block of uniaxially compacted MX-80 clay powder using 100 MPa pressure. The

diameter is 1.7 m and the height 0.5 m. The water content is about 12% and the density

approximately 2000 kg/m

Chapter 11.4: Clays and Nuclear Waste Management708

B. Uptake of Water from the Rock (THMC)

Uptake of water from the rock and backﬁll is the process that ultimately yields

complete water saturation and the ﬁnal density distribution. If the rock gives off very

little water to the clay, the backﬁll in the overlying tunnel may serve as the major

water source. The water transport from the backﬁll to the buffer is controlled by the

water pressure and by the suction of the buffer, as well as by the degree of water

saturation of the backﬁll.

The wetting has a (C)-effect since Ca

,Mg

,SO

2

and Cl



ions will be

precipitated at the wetting front which moves towards the canister. Thus, a rather

large part of the buffer may contain such precipitates, which interact with the

smectite minerals and the canisters. The temperature gradien t has also a (C)-effect.

This is because silicate minerals (including the smectite component) that are close to

the heater will dissolve to a larger extent than those at the rock. The mechanism by

which released silica migrates towards the rock and precipitates in the outer part of

the bentonite is not well understood. Siliciﬁcation causes some cementation and

brittleness of the clay.

11.4.3. NUMERICAL MODELLING

The evolution of a detailed design of the buffer (the interacting components for

the KBS-3 concept are deﬁned in Fig. 11.4.5) has to be modelled with respect to

maturation and homogenisation. Current co-operative international work, partly

supported by the EU, deals with these issues. The conceptual model is the basis for

the selection and development of numerical tools to predict and evaluate the T, H, M

and C processes in the engineered barriers.

The proposed numerical formulations are exempliﬁed by the EU-supported on-

going work ‘‘Prototype Repository Project’’ by H.R Thomas and P.J Cleall (per-

sonal communication) at Cardiff University. The main features of their numerical

model, code-named COMPASS, are

1. Moisture ﬂow involves both liquid ﬂow and vapour transfer. Liquid ﬂow is as-

sumed to follow a generalised Darcy’s law, while vapour transfer is represented by

a modiﬁed Philip and de Vries approach.

2. Heat transfer includes conduction, convection and latent heat of vaporisation

transfer in the vapour phase.

3. Flow of dry air due to the bulk ﬂow of air and that of dissolved air in the liquid

phase are considered. The bulk ﬂow of air, induced by an air pressure gradient, is

again represented by a generalised Darcy’s law. Henry’s law is employed to cal-

culate the quantity of dissolved air and its ﬂow is coupled to the ﬂow of pore-liquid.

4. Deformation effects are included via either a non-linear elastic approach or an

elastoplastic formulation. In both cases de formation is taken to be dependent on

suction, stress and temperature changes.

11.4.3. Numerical Modelling 709

5. Chemical solute transport for multi-chemical species includes diffusion, dispersion

and accumulation from reactions due to sorpti on.

The basis for formulation of governing equations is that heat conduction and ﬂow

are expressed using classical physics but is generalised by including the velocities of

liquid, vapour and air. The velocities of pore-liquid and pore-air are calculated using

a generalised Darcy’s law with special respect to the chemical solute concentration

gradient, the conductivity of the air phase and the pore-air pressure. Also, an

osmotic ﬂow term in the liquid velocity is included to represent liquid ﬂow behaviour

in some highly compacted clays. Air in partly saturated soil may exist in two forms:

bulk air and dissolved air. In this approach, the proportion of dry air in the pore-

liquid is deﬁned using Henry’s law.

Fig. 11.4.5. Detail of KBS-3 deposition hole with copper-shielded canister embedded in

compacted buffer blocks and bentonite pellet ﬁlling.

Chapter 11.4: Clays and Nuclear Waste Management710

When a chemical solute is non-reactive, its sorption on to the soil surface may be

ignored. The governing equation for chemical transfer can then be expressed in terms

of diffusion and dispersion. This equation, given in primary variable form, can

be extended to a multi-chemical species form by introducing a sink term to account

for mass accumulation from reactions due to sorption. This is then coupled to a

geochemical model.

The total strain consists of components related to suction, temperature, chemical

and stress changes. Formulations have been employed, taking into account both

elasticity and elastoplasticity. The contribution of chemical solute to the stress-strain

behaviour of the soil can, to a ﬁrst approximation, be described by an elastic state

surface concept based on osmotic potential.

A numerical solution of the governing differential equations is achieved by using

the Galerkin weighted residual method to formulate the ﬁnite element discretisation.

Shape functions are used to deﬁne approximation polynomials in the ﬂow and stress/

strain equations. The outcome of predictive modelling for the evolut ion of the ‘‘nea r-

ﬁeld’’ KBS-3 concept, made at Cardiff University, are shown in Figs. 11.4.6 (basic

geometry), 11.4.7 (pore-water pressure in the buffer in contact with the canister) and

11.4.8 (temperature at the canister/ buffer contact).

Both the theoretical model and the full-scale experiments actually show in-ﬂow

from the host rock overwhelming the predicted drying close to the hot canister,

leading to eventual saturation (Fig. 11.4.7). The peak in temperature evolution

(Fig. 11.4.8) represents the state when the buffer is at its driest. This demonstrates

the effect of coupling between the hydraulic ﬁeld and the temperature ﬁeld, and that

57.8

Fig. 11.4.6. Tunnel and rock section used for the thermal analysis and enlarged view of cross-

section (ongoing modelling at Cardiff University for the EU-supported project ‘‘Prototype

Repository Project’’).

11.4.3. Numerical Modelling 711

0.5 10

024 810

-0.5 10

-1 10

-1.5 10

-2 10

-2.5 10

-3 10

135

Time (Years)

67 9

Pore Water Pressure (Pa)

Fig. 11.4.7. Pore-water pressure plot at the heater/buffer interface at mid-heater height (on-

going modelling at Cardiff University for the EU-supported project ‘‘Prototype Repository

Project’’).

100

05678910

Time (Years)

1234

Temperature (°C)

Fig. 11.4.8. Temperature plot at the heater/buffer interface at mid-height of the heater (on-

going modelling at Cardiff University for the EU-supported project ‘‘Prototype Repository

Project’’).

Chapter 11.4: Clays and Nuclear Waste Management712

relevant numerical predictions should be based on coupled (THMC) models when

possible.

11.4.4. LONG-TERM PHYSICAL PERFORMANCE OF THE BUFFER

For physical stability of the buffer clay two major factors are identiﬁed: colloid

transport and ﬁltering, and microbiological ﬁltering.

The issue of colloid transport through buffers was not examined in detail but

some general conditions were deﬁned. For example, by sorbing pos itively charged

radionuclides and other cations, very small smectite particles can transport con-

taminants to the biosphere. Like many other constituents of colloidal size, such

particles can form aggregates, gels and networks that easily link to the dense buffer

clay (Pusch, 1999). Although smectite aggregates can be released if the pore-water

ﬂow is sufﬁciently high, this seems impossible under the local hydraulic gradients

that prevail in the rock after closing the repository.

However, the buffer clay will tend to enter fractures in the rock although the front

of fracture-penetrating clay may be very soft. Groundwater ﬂowing in the fractures

may, therefore, tear off small aggregates and transport them if the ﬂow rate exceeds a

few millimeters per second. Experiments as well as ﬁeld observat ions demonstrate

that the size of aggregates may range from 2 to 50 mm. Because of their surface

charge the aggregates can interact with, and attach to, the common fracture min-

erals, such as chlorite and micas. For this reason, the transport of smectite aggre-

gates over long distances is not significant.

Bacteria may stay alive and develop in soft smectitic clay. Those that reduce sulphate

are of particular importance to engineered barriers of the SKB-type as they usually

carry a negative surface charge, and hence can move through clay like large organic

polymers, such as humus colloids. However, such movement can be limited, at least

partly, by hydration conditions and the very small pore space that exists in dense buffer

clay (Pedersen, 1997). For MX-80 type clay with a density higher than 1800 kg/m

,the

small size of interconnected voids and the limited access to free water prevent bacteria

from moving and staying alive. Although spores may survive and be operative, the

working hypothesis is that ultimately all life forms in the buffer will be eradicated.

11.4.5. CHEMICAL STABILITY OF THE BUFFER

While gamma radiation has a small impact on the longevity of the buffer smectites

(Pusch, 1994), dissolution and conversion to non-expanding clay minerals is a threat.

The chemical integrity of smectites is determined by thermodynamically controlled

reactions. Thus, mineral stability and transformations depend on the prevailing

conditions of temperature and temperature gradient under which minerals and pore-

water interact.

11.4.4. Long-Term Physical Performance of the Buffer 713