atw 2018-05v6

atw Vol. 63 (2018) | Issue 5 ı May
(a) Parameter variation: t cool
(b) Parameter variation: k
(c) Parameter variation: D 1
50 u(t,P) [°C]
40
30
20
10
0
0 1 10 10 2 10 3 10 4 10 5
50 u(t,P) [°C]
40
30
20
10
0
55 y
100 y
200 y
300 y
450 y
(d) Parameter variation: D 2
0 1 10 10 2 10 3 10 4 10 5
50 u(t,P) [°C]
40
30
20
10
0
7.6 m
9 m
11 m
15 m
asymp.
(g) Parameter variation: n, D 1 , D 2
27, 40 m, 7.6 m
9, 65 m, 12 m
9, 65 m, 15 m
1, −, 15 m
asymp.
0 1 10 10 2 10 3 10 4 10 5
| | Fig. 3.
Time [y]
Temperature evolution at P = P 2 for the
baseline configuration C a (black) and
alternative configurations (blue).
A-F: One parameter changed (parameter in
title); G: Three parameters changed.
Host rock composition as in Figure 2.
peak for D 1 = 40 m, this is reduced to
just 32 y to peak for D 1 = 60 m. The
envelope area required for siting the
repository increases proportionally
to drift spacing. This option requires
to increase the length of the access
tunnel to the drifts, while the total
length of the disposal drifts remains
unchanged. An increase from 40 m to
60 m increases space requirements by
50 % in envelope area. The cost and
side effects have to be weighed against
the benefit of a significantly lower and
earlier temperature peak. Also, with
more space available between drifts,
retrieval could be facilitated.
Increasing batch spacing D 2 from
7.6 m to 15 m leads to a remarkable
reduction of peak temperature at P 1 ,
cutting the amplitude by over 40 %
(Figure 2d, Table 3a). Peak time
remains roughly constant for small
D 2 and decreases abruptly between
11 m and 15 m. This option requires
doubling the total length of the disposal
drifts, while the number of
batches remains unchanged. The area
for disposal increases proportionally
50 u(t,P) [°C]
40
30
20
10
0
0 1 10 10 2 10 3 10 4 10 5
50 u(t,P) [°C]
40
30
20
10
0
k = 4
k = 3
k = 2
k = 1
(e) Parameter variation: n
0 1 10 10 2 10 3 10 4 10 5
Time [y]
n = 27
n = 7
n = 3
n = 1
to batch spacing. This cost has to be
weighed against a significant reduction
of peak temperature. Retrieval is
handicapped on the one side by longer
drifts and favoured on the other by
substantially decreased temperatures.
Passing from 27 to 3 disposal drifts
leads to a limited reduction of peak
temperature, while passing from 3
to 1 drift leads to a more significant
drop (Figure 2e). Accordingly, the
reduction in peak temperature per
row is slight for large n and important
for small n (Table 3a). Similar to the
effect of drift spacing, the temperature
peak shortens significantly. For
n = 3, the temperature peaks 40 y
after waste disposal (compared with
316 y in base line configuration).
Decreasing the numbers of drifts
significantly influences the geometry
of the repository; the area required
remains constant, but its envelope
becomes long and narrow. For low
values of n, the difficulty of “fitting”
the repository into a host formation
may increase. In panel e, the number
of drifts is reduced in steps from 27 to
1. Accordingly, their length increases
from 616 m (27 drifts) to 2.4 km
(7 drifts) to 5.6 km (3 drifts) and to
16.7 km for a single drift. The technical
effort consists in ensuring longer
operating times for the individual
drifts.
Finally, the repository can be
divided into clusters (in some cases, it
might have to, due to the presence of
determining geological features). The
clusters are assumed to be sited in
sufficient distance to remain thermally
unaffected from each other. The
effect is shown in Figure 2f. Thermal
benefits on peak temperature are
limited unless the clusters are made
50 u(t,P) [°C]
40
30
20
10
0
0 1 10 10 2 10 3 10 4 10 5
50 u(t,P) [°C]
40
30
20
10
0
(f) Parameter variation: s
Time [y]
40 m
50 m
60 m
80 m
asymp.
27 x 81
15 x 45
9 x 27
3 x 9
1 x 1
0 1 10 10 2 10 3 10 4 10 5
very small (i.e. holding well below
10 % of the total number of batches).
The required area per cluster decreases
proportionally to the number
of batches it holds, which can be an
advantage for “fitting” the repository
in perturbed sites. In panel f, the
disposal area is fractioned into clusters
of various size (indicated by n×m,
with s = nm| option / nm| baseline ). As
for panel c, this option requires to
increase the length of the access
tunnels.
Panel g presents results for
alter native configurations with n
decreased and D 1 , D 2 increased, other
parameters unchanged. The asymptotic
case is represented by the configuration
D 1 , D 2 → ∞. The hatched area
represents the scope for thermal
dimensioning below the temperature
curve belonging to the baseline configuration.
A large fraction of the
scope for thermal dimensioning is
available if triple the amount of space
with respect to the baseline configuration
is reserved, allowing for efficient
action on D 1 , D 2 and n.
Another point of interest is the
thermal benefit at the time horizon
10 4 y, when steel waste canisters are
expected to fail. At this point, the
ambient conditions in the canister core
change from dry to wet, thus enabling
liquid-phase reactions that were previously
excluded. For thermally activated
diffusion-reaction processes, the
rates depend on the ambient temperature.
While all engineering parameters
can be used to substantially reduce
the temperatures in first 10 3 y, this is
no longer the case in the following 10 4
y. Except for the asymptotic cases,
there is little response in temperatures
at the horizon 10 4 y for changes in t cool ,
DECOMMISSIONING AND WASTE MANAGEMENT 323
Decommissioning and Waste Management
Scope for Thermal Dimensioning of Disposal Facilities for High-level Radioactive Waste and Spent Fuel ı Joachim Heierli, Helmut Hirsch, Bruno Baltes