1536
Proceedings of the 18
th
International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
were carried out at loading frequency in the range from about
0.01 Hz to 0.2 Hz.
Menq (2003) used both an RCTS device and an MMD
(Multi-Mode Resonant) device to investigate the dynamic
properties of gravelly soils. The RCTS device is capable of
performing on the same soil specimen both the torsional
resonant column (RC) test at high loading frequency (i.e., the
resonant frequency) and in the nonlinear range and the cyclic
torsional shear (TS) test at much lower frequencies, simply by
changing the amplitude and frequency of the current in the drive
coils and the motion monitoring devices used to record the
specimen response (Isenhower 1979; Ni 1987; Hwang 1997).
Because the same specimen can be subjected to both the RC and
TS tests, it eliminates the variability due to testing different
specimens or testing the same specimen subjected to a different
stress history (Darendeli 2001). The test specimen for RC or TS
testing typically has a diameter in range from 36 to 76 mm and
a height from 72 to 152 mm (Menq and Stokoe 2003). To
accommodate gravelly specimens with relatively large particle
sizes, the MMD was developed and is capable of testing
specimen with 152 mm in diameter and 600 mm in height in
different measurement modes, including the torsional resonance
mode similar to resonant column tests.
Based on the test results, Menq (2003) used the modified
hyperbolic model suggested by Darendeli (2001) to model for
shear modulus reduction of gravelly soils:
]) / ( 1[1
/
max
a
r
GG
(1)
5.0
5.0
0
6.0
) /
(
12.0
u
C
a
u
r
p
C
(2)
) /
log(
1.0 86.0
0
a
p
a
(3)
where the reference strain
r
(%) is
at G/G
max
= 0.5,
a
is the
curvature coefficient,
C
u
is the uniformity coefficient,
0
’
is the
effective isotropic confining pressure, and
p
a
is the atmospheric
pressure (1 atm).
3 MECHANICAL PROPERTIES OF TESTED GRAVEL
Two types of gravel were tested for potential use as engineered
fill in this study. They are aggreagates derived from processing
crushed stone mined from a rock quarry, consisting of angular
and hard particles, with one of them being poorly graded and
designated as PA and the other being relatively well graded and
designated as WA. Three batches of the WA material (WA-1,
WA-2, and WA-3) and one batch of the PA material (PA-1)
were taken for testing.
Modified Proctor tests in accordance with ASTM D1557
were performed on the WA material (WA-1 and WA-3) after
removing/scalping particles greater than 19 mm in diameter.
The modified Proctor test is not applicable to the PA material
according to ASTM 1157. To be consistent with the modified
Proctor test, all the other laboratory tests were also performed
on the scalped material. Figure 2 shows the typical grain size
distribution curves for the tested materials (i.e., PA and WA), as
well as the grain size distribution curves of each batch of the
material after scalping particles greater than 19 mm in diameter
(i.e., PA-1, WA-1, WA-2, and WA-3).
In addition, maximum and minimum index densities were
obtained based on ASTM 4254 and ASTM 4253 for both the
PA material (PA-1) and the WA material (WA-1 and WA-3).
As seen in Table 1, the maximum index density of the WA
material determined using a vibratory table is very close to the
maximum density obtained by impact compaction in which the
moisture-density relationship is defined. But comparison shows
that the maximum index density of the WA material is
significantly (about 40%) higher than that of the PA material,
which is understandable as the voids between the larger
0
10
20
30
40
50
60
70
80
90
100
0.001
0.01
0.1
1
10
100
Grain Size (mm)
Percent Finer by Weight (%)
PA-1 (Scapled)
WA-1 (Scapled)
WA-2 (Scapled)
WA-3 (Scapled)
Typical
PA
Typical WA
Gravel
Sand
Fines
Figure 2. Grain size distribution curves of the unscalped and scalped
(tested) gravel specimens.
particles of the WA material are filled with smaller particles.
From each batch of the material, a pair of specimens
designated as A and B were created for the RCTS tests. The pair
of specimens from each batch of the WA material were
separately remolded at the optimum moisture content to
approximately 95% and 100% of the maximum dry density
determined in the modified Proctor test. And the pair of
specimens of the PA material were remolded to relative
densities of about 80% and 100% at a moisture content of about
1%. All specimens were compacted to the target densities using
a hammer drill fitted with a specifically designed circular steel
face of 146 mm in diameter.
After the RCTS tests were completed, more index tests such
as the water content and dry density were performed on each
specimen, and the results are summarized in Table 2, including
the derived degrees of saturation and void ratio.
Table 1. Mechanical properties of scalped gravel samples.
D
50
(mm)
C
u
C
c
min
max
e
min
e
max
max
(Mg/m
3
)
w
opt
(%)
PA-1 11.8 2.1 1.3 2.83 1.39 1.64 0.73 1.04 -
-
WA-1 3.4 174.5 3.81 2.72 1.67 2.30 0.18 0.62 2.31 0.653
WA-2 -
-
-
2.72 -
-
-
-
2.30 0.469
WA-3 3.2 150.6 4.87 2.82 1.67 2.27 0.24 0.69 2.34 0.653
Index Void
Ratio
Moisture-
Density
Relationship
Sample
Name
Grain Size
Distribution
G
s
Index
Density
(Mg/m
3
)
Note: D
50
is the particle diameter corresponding to 50% passing; C
c
is
the coefficient of curvature, G
s
is the specific gravity, and w
opt
is the
optimum moisture content.
Table 2. Mechanical properties of gravel specimens tested in the RCTS
device.
Sample
Name Specimen
Water
Content
(%)
Saturation
(%)
Dry
Density
(Mg/m
3
)
Void
Ratio
A
1
3.5
1.57 0.81
B
0.8
3.2
1.66 0.70
A
6.4
72.1
2.19 0.24
B
6.1
85.5
2.27 0.19
A
5.5
59.8
2.17 0.25
B
4.4
65.3
2.30 0.18
A
5.8
61.5
2.23 0.27
B
6.2
87.2
2.35 0.20
PA-1
WA-1
WA-2
WA-3
4 RCTS TESTS ON COMPACTED GRAVEL
During RCTS testing, the specimen is sealed in a membrane,
and the pore pressure in the specimen is vented to atmosphere
pressure. From the results of cyclic triaxial tests on Toyoura
sand, Kokusho (1980) indicated that the drained tests and the
undrained tests give almost identical strain-dependent variation
of the modulus within the strain level from 10
-4
% to 0.5%.
Since the gravel specimens have larger permeability due to the
