330
Proceedings of the 18
th
International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
Proceedings of the 18
th
International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
2
0,01
0,1
1
10
100
0
20
40
60
80
100
Cut grading
(CG)
CG
D
max
= 10 mm
D
50
= 2.92 mm
U
c
= 19.0
F
c
= 9.23 %
Parallel grading
(PG)
PG
D
max
= 10 mm
D
50
= 0.72 mm
U
c
= 50.0
F
c
= 16.5 %
OKG
D
max
= 37.5 mm
D
50
= 6.21 mm
U
c
= 44
F
c
= 4 %
SCG/SAG
D
max
= 10 mm
D
50
= 2.32 mm
U
c
= 19.0
F
c
= 5.2 %
OBG
D
max
= 38.1 mm
D
50
= 2.92 mm
U
c
= 50.0
F
c
= 9.23 %
Sieved grading
(SCG/SAG)
Original crushed
sandstone JR (OBG)
Percent finer than D by weight
Particle size, D (mm)
Original crushed
sandstone Kikonai (OKG)
Figure 1. Grading curves and characteristics of the tested cement-mixed
gravels
a)
5
6
7
8
9
10
11
2.0
2.1
2.2
Test conditions (w
opt
;D
c
=95%)
1.0 Ec
c/g=2.5%
4.5 Ec
c/g=4/0%
Dry density,
d
(g/cm
3
)
Sieved Chiba gravel (SCG)
water content, w (%)
b)
4
5
6
7
8
9 10
2.1
2.2
2.3
Modified Proctor (4.5E
c
)
Test conditions (w
opt
;D
c
=95%)
7.9 5.4 6.9 7.3 7.75
w
opt
(%):
p
d_max
(
g/cm
3
)
:
water content, w (%)
Dry density,
d
(g/cm
3
)
OBG
PG
CG
OBG OKG CG PG SAG
2.22 2.23 2.23 2.21 2.22
SAG
OKG
Figure. 2. Compaction curves: a) SCG for 1.0Ec & 4.5 E
c
and c/g= 2.5%
& 4/0 %; and b) five cement-mixed gravelly soils presented in Fig. 1 for
4.5Ec and c/g= 4.0% (except for PG and OKG)
Fig. 3 presents the development of the compressive strength
q
max
at confining pressure of 20 kPa with the period of initial curing
under the atmospheric pressure at a constant water content (the
same as prepared) of four kinds of SCG specimen prepared under
different conditions (1.0E
c
or 4.5E
c
and
c/g
=2.5 or 4.0 %). The
following trends can be noted. Firstly, the
q
max
value increases
considerably with time, which should be due to the cement
hydration process. The increase until a curing period of 14 days
is rather proportional to the “initial” value at 7 days.
Secondly,
the
q
max
value is largely different among the four kinds of
specimen (up to a factor of 100 %). Thirdly, the effects of
compacted dry density
ρ
d
on the
q
max
value when
c/g
= 2.5 % are
significant. An increase more than 100 % results from an
increase in
ρ
d
of only about 5 % associated with an increase in
the compaction level from 1E
c
to 4.5E
c
. On the other hand, when
the compaction level is 1.0E
c
, the
q
max
value increases by a factor
of only about 40 % with an increase in
c/g
from 2.5 % to 4. 0%
(i.e., an increase of about 60 %).
E
c
and
c/g
are the parameters commonly used in practice,
because they are easily measured and controlled. However, they
are not the basic parameters that control the strength and
deformation characteristics of CMG. This feature can be easily
seen from the following inherent drawbacks (Watanabe et al.
2003, Kongsukprasert et al. 2005). Firstly, with materials having
different specific density
s
, for the same c/g value and the same
soil void ratio, the volume of cement per volume increases with
an increase in
s
, despite that
s
has no direct effect on the stress-
strain properties. Secondly, the effect of compaction level for the
same
c/g
value (and same
s
) has two components: 1) a better
interlocking among soil particles with a decrease in the soil void
ratio; and 2) a larger amount of cement in a less volume of the
total void of soil particle skeleton. Based on this consideration,
two independent parameters are postulated: the soil skeleton
porosity
n
s
(representing the structure of the skeleton of gravelly
soil particles only); and the cement void ratio
C
r
(representing
the fraction of the void of the soil skeleton occupied by cement):
s
s
V V n
V
(1)
c
c
r
v
s
V V
C
V V V
(2)
where
V
is the total volume;
V
s
is the volume of gravelly soil
particles;
V
v
is the volume of the void of the skeleton of gravelly
soil ; and
V
c
is the volume of cement.
Ezaoui et al. (2011) also proposed the following hyperbolic
function for the
q
max
value that increases with time and is a
function of these two parameters independently:
max
0
.
1
r c
c
s
c
a C t
q t
q n
b t
(3)
where
t
c
is the curing period;
q
0
is the initial compressive
strength (when
t
c
= 0) that decreases with
n
s
;
a
is the parameter
showing the cementation effect that increases with
C
r
; and
b
is
the constant parameter that depends on cement type. The
functions
a
(
C
r
) and
q
0
(
n
s
) are obtained based on the data
presented in Fig. 3 together with those from CD TC tests on
rotary core samples retrieved from the field, as shown below.
Three solid lines presented in Fig. 4 denote the iso-strength lines
for constant
q
max
=
q
1
,
q
2
and
q
3
at a specified
t
c
according to Eq.
(3) with known values of
q
0
,
a
and
b
. The dash-dot curves are the
corresponding
c/g
= constant curves. From such a plot as shown
in Fig. 4, the most suitable (i.e., the most cost-effective)
combination of (
C
r
;
n
s
) of a given type of CMG that achieves a
given required compressive strength can be chosen referring to
the cost for cement and compaction work.
2.2
U tests on specimens made using a material use
in the field
The use of CMG is now spreading in Japan, particularly to
construct bridge abutments for high speed train lines. Very
recently, a geosynthetic-reinforced soil (GRS) integral bridge
was constructed. The backfill immediately behind the facing is
well-compacted cement-mixed gravelly soil. The grading curve
of this backfill material (i.e., crushed gravel from a quarry,
denoted as the Original Kikonai Gravel, OKG) is presented in
Fig. 1. The compaction curve for 4.5E
c
is presented in Fig. 2b.
Before the construction, to determine the cement-mixing
proportion, a series of U tests were performed on specimens
mixed at c/g = 2.0, 4.0 and 6.0 %, compacted to
D
c
=
95 % at
w=
w
opt
(4.5E
c
) and cured for 7 and 28 days. To accommodate the
maximum particle size of 37.5 mm, large specimens (150 mm in
diameter and 300 mm in height) have been used.
To apply Eq. (3) to the data from these U tests,
b
=19.62 was
used, which was obtained by analyzing a data set from a series of
CD TC tests on cement-mixed SCG cured for a period of 1 ~ 180
days (Ezaoui et al., 2010, 2011). The values of
a
(
C
r
) and
q
0
(
n
s
)
for the three
c/g
values and the two curing periods were
determined from the reported values of
q
max
(
t
c
). They are plotted
against their corresponding
n
s
and
C
r
values in Figs. 6a and b
(round symbols). In so doing, to eliminate possible effects of
specimen size and grading characteristics (discussed in the next
section), these
q
max
values had been corrected to become the
same values as those obtained for the smaller specimen (i.e., the
data presented in Fig. 3) under the same test conditions (i.e., a
curing period of 7 days;
D
c
= 95 %; and
c/g
= 4.0 %). Then, in
Figs. 6a and b, only relative variations of
q
0
and the parameter
a
due to variations in the
n
s
and
C
r
values among those U test
