329
Proceedings of the 18
th
International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
1
Strength properties of densely compacted cement-mixed gravelly soil
Ppropriétés de résistance des graves cimentées fortement compactées
Ezaoui A.
CETE de Lyon, DLA, France
Tatsuoka F., Furusawa S., Hirao K., Kataoka T.
Tokyo University of Science, Civil Engineering Department, Japan
ABSTRACT: A series of drained triaxial compression tests (TC) were performed on specimens prepared in the laboratory and rotary
core samples retrieved from the field of well-compacted cement-mixed gravelly soils (CMG). The effects of the following factors on
the compressive strength
q
max
were evaluated: the degree of compaction
D
c
and cement content (when compacted at the optimum
water content); the curing time; the grading characteristics of gravelly soil; and the specimen volume. The followings were found
from the data for a wide variety of CMG having
c/g=
2.0 ~ 6.0 % compacted by using two energy levels (1.0Ec & 4.5Ec). The effects
of
D
c
are equally important as the cement content. An empirical equation is proposed using two independent variables, the soil
skeleton porosity
n
s
, which controls the initial compressive strength, and the cement void ratio
C
r
, which controls the increasing
manner of
q
max
with curing time. The effects of sieving method to obtain smaller particle materials for small TC specimens from an
original gravelly soil having larger particles are not significant. However, significant effects of specimen volume (i.e., 72 mm x 72
mm in cross-section times 150 mm high versus 300 mm in diameter times 580 mm high) is highlighted.
RÉSUMÉ :
Une campagne d’essais triaxiaux en compression a été réalisée sur des échantillons de grave, cimentée et compactée
(GC), préparés en laboratoire mais également sur des échantillons carottés sur site. Les effets de différents facteurs sur les résistances
en compression
q
max
ont été évalués : le degré de compaction
D
c
(teneur en eau de l’optimum Proctor), la teneur en ciment, le temps de
cure; la distribution granulométrique et le volume de l’échantillon. Les tendances suivantes ont pu
être relevées à partir de données
expérimentales obtenues sur une grande variété de GC présentant différentes teneurs en ciment
c/g=
2.0 ~ 6.0 % et compactées
suivant deux modalités: normale et modifiée (1.0Ec & 4.5Ec). Il est apparu que les effets de la compaction ou de la teneur en ciment
sur la résistance en compression étaient tout aussi importants. Une formulation empirique est proposée afin d’évaluer la résistance en
compression à partir de deux variables indépendantes : la porosité du squelette
n
s
, qui contrôle la résistance initiale en compression, et
la proportion du volume des vides occupée par le ciment
C
r
, qui contrôle l’augmentation de
q
max
avec le temps de cure. Les effets
d’une réduction granulométrique ne sont pas significatifs. Cependant, les effets du volume de l’échantillon sont
mis en évidence.
KEYWORDS: cement-mixed gravel, curing, triaxial compression, compaction, size effect
1 INTRODUCTION
Ground improvement by cement-mixing has been successfully
used in many construction projects. This technology includes
mixing-in-place of soft clay without compaction (i.e., the deep
mixing method), under-water placing of cement-mixed soil slurry
without compaction and highly compacted cement-mixed gravel
for high roller-compacted concrete (RCC) dams. More recently,
cement-mixed gravelly soil (CMG) compacted by energy lower
than the one for RCC is used to construct bridge abutments for
high-speed trains in Japan. Such use of CMG for deformation-
sensitive structures as above has been motivated by a high cost-
effectiveness. To develop the design and construction
procedures, the stress-strain properties of CMG have been
studied by many researchers (Lohani et al. 2004, Kongsuprasert
& Tatsuoka 2005; Kongsukprasert et al. 2005, 2007; Tatsuoka et
al. 2008, Ezaoui et al. 2010). Yet, the whole picture of effects of
dry density, water content at compaction and cement content
ratio on the strength and stiffness are not well understood (e.g.,
Horpibuslsuk et al. 2003; Consoli et al. 2007).
One of other practically important issues is the use of well-
graded gravelly soil in the field with the maximum diameter
exceeding 35 mm, which is too large to be tested by ordinary
laboratory stress-strain tests, such as triaxial compression (TC)
tests. For this reason, it is usually prepare materials for
specimens for TC tests, for example, by sieving out large
particles from a given original gravelly soil. So, it becomes
necessary to evaluate the effects of grading characteristics (i.e.,
the particle size and the shape of grading curve) and the volume
of specimen.
The present study aims at finding major factors that control the
strength and stiffness of CMG for its proper and cost-effective
use. In particular, to evaluate the effects of compacted degree of
compaction, cement content, curing time, particle grading
characteristics and specimen volume were evaluated, a
comprehensive series of drained TC tests were performed.
2 EFFECTS OF SEVERAL FACTORS ON STRENGTH
2.1
Specimens prepared in the laboratory
The first TC series performed in the present study used
SievedChiba Gravel (SCG in Fig. 1, crushed quarry sandstone
comprising sub-angular particles). The TC specimens were
rectangular prismatic (72 mm x 72 mm in cross-section times
150 mm high), for which a maximum particle size of 10 mm was
selected. Fig. 2a shows compaction curves for standard and
modified Proctor energy levels (1.0Ec= 550 kJ/m
3
and 4.5E
c
) and
cement/gravel ratio in weight
c/g
= 2.5 % and 4.0 % of SCG.
d_max
=2.12 g/cm
3
and
w
opt
= 9.3 % were obtained for 1Ec and
d_max
=2.21 g/cm
3
and
w
opt
= 7.4 % for 4.5Ec. Effects of cement-
mixing on the compaction curves are negligible (Ezaoui et al.,
2010). The TC specimens were produced by tamping moist
cement-mixed gravelly soil immediately after adding water for
respective optimum water contents
w
opt
to a degree of
compaction
D
c
= 95 %, where
D
c
=
d_test
/
d_
max
x 100 % for 1.0E
c
or 4.5E
c
(Fig. 2a). The weight and height of each of the five sub-
layers were carefully controlled. In the test results shown
hereafter, vertical (axial) and horizontal (lateral) strains,
v
and
h
, measured locally with of a pair of vertical local deformation
transducers (LDTs) arranged on two opposite lateral faces of
specimen and horizontal LDTs arranged on the other two
opposite lateral faces are presented.
tr
t
r
rti
f
l
t
t- i
r ll
il
ropriétés de résistance des graves cimentées fortement compactées
Ezaoui A.
CETE de Lyon, DLA, France
Tatsuoka F., Furusawa S., Hirao K., Kataoka T.
Tokyo University of Science, Civil Engineering Depart ent, Japan
: series of drained triaxial co pression tests ( ) ere perfor ed on speci ens prepared in the laboratory and rotary
core sa ples retrieved fro the field of ell-co pacted ce ent- ixed gravelly soils (
). he effects of the follo ing factors on
the co pressive strength
q
max
ere evaluated: the degree of co paction
c
and ce ent content ( hen co pacted at the opti u
ater content); the curing ti e; the grading characteristics of gravelly soil; and the speci en volu e. he follo ings ere found
fro the data for a ide variety of
having
c/g
2.0 6.0 co pacted by using t o energy levels (1.0 c 4.5 c). he effects
of
c
are equally i portant as the ce ent content. n e pirical equation is proposed using t o independent variables, the soil
skeleton porosity
n
s
, hich controls the initial co pressive strength, and the ce ent void ratio
r
, hich controls the increasing
anner of
q
ax
ith curing ti e. he effects of sieving ethod to obtain s aller particle aterials for s all
speci ens fro an
original gravelly soil having larger particles are not significant. o ever, significant effects of speci en volu e (i.e., 72 x 72
in cross-section ti es 150 high versus 300 in dia eter ti es 580 high) is highlighted.
:
ne ca pagne d’essais triaxiaux en co pression a été réalisée sur des échantillons de grave, ci entée et co pactée
( ), préparés en laboratoire ais égale ent sur des échantillons carottés sur site. es effets de différents facteurs sur les résistances
en co pression
q
max
ont été évalués : le degré de co paction
c
(teneur en eau de l’opti u roctor), la teneur en ci ent, le te ps de
cure; la distribution granulo étrique et le volu e de l’échantillon. es tendances suivantes ont pu ê
tre relevées à partir de données
expéri entales obtenues sur une grande variété de présentant différentes teneurs en ci ent
c/g
2,0 6,0 et co pactées
suivant deux odalités: nor ale et odifiée (1,0 c et 4,5 c). Il est apparu que les effets de la co paction ou de la teneur en ci ent
sur la résistance en co pression étaient tout aussi i portants. ne for ulation e pirique est proposée afin d’évaluer la résistance en
co pression à partir de deux variables indépendantes : la porosité du squelette
n
s
, qui contrôle la résistance initiale en co pression, et
la proportion du volu e des vides occupée par le ci ent
r
, qui contrôle l’aug entation de
q
max
avec le te ps de cure. es effets
d’une réduction granulo étrique ne sont pas significatifs. ependant, les effets du volu e de l’échantillon sont is en évidence.
: ce ent- ixed gravel, curing, triaxial co pression, co paction, size effect
