Professional Documents
Culture Documents
Stiffness of Sands Through A Laboratory Test Database
Stiffness of Sands Through A Laboratory Test Database
078]
Deformations of sandy soils around geotechnical structures generally involve strains in the range small
(0.01%) to medium (0.5%). In this strain range the soil exhibits non-linear stress–strain behaviour,
which should be incorporated in any deformation analysis. In order to capture the possible variability
in the non-linear behaviour of various sands, a database was constructed including the secant shear
modulus degradation curves of 454 tests from the literature. By obtaining a unique S-shaped curve of
shear modulus degradation, a modified hyperbolic relationship was fitted. The three curve-fitting
parameters are: an elastic threshold strain ªe , up to which the elastic shear modulus is effectively
constant at G0 ; a reference strain ªr , defined as the shear strain at which the secant modulus has
reduced to 0.5G0 ; and a curvature parameter a, which controls the rate of modulus reduction. The two
characteristic strains ªe and ªr were found to vary with sand type (i.e. uniformity coefficient), soil
state (i.e. void ratio, relative density) and mean effective stress. The new empirical expression for
shear modulus reduction G/G0 is shown to make predictions that are accurate within a factor of 1.13
for one standard deviation of random error, as determined from 3860 data points. The initial elastic
shear modulus, G0 , should always be measured if possible, but a new empirical relation is shown to
provide estimates within a factor of 1.6 for one standard deviation of random error, as determined
from 379 tests. The new expressions for non-linear deformation are easy to apply in practice, and
should be useful in the analysis of geotechnical structures under static loading.
54
STIFFNESS OF SANDS THROUGH A LABORATORY TEST DATABASE 55
s9 ¼ ( 19 þ 39 )=2 rather than p9 ¼ ( 19 þ 29 þ 39 )=3: new database has been constructed incorporating shear mod-
Nevertheless, it will be p9 that will be used in the ulus degradation curves from the literature. This curve-fitting
following, chiefly because that is what the original authors process has led to new interpretations and definitions that
invariably quoted. Once recognised, of course, this con- facilitate prediction of the shear modulus degradation of
stant arithmetical factor can easily be accounted for in sands with strain, based on elementary soil classification
plane-strain applications, for example. data.
Seed & Idriss (1970) published the first database of shear
modulus degradation curves for sand, for the purpose of
earthquake site response analysis. This S-shaped curve was STIFFNESS DATABASE OF SANDY SOILS
obtained for 75 tests on a total of 30 sands, with a wide Because of the difficulty of sampling, and the require-
range of confining pressure, relative density and void ratio. ments of sophisticated element tests, most engineers prefer
Hardin & Drnevich (1972b), Iwasaki et al. (1978), Kokusho in situ testing to laboratory testing to determine the stiffness
(1980) and the many others showed that equation (1) can be of sands. However, extensive laboratory stiffness data al-
used to calculate a shear modulus for small to medium ready exist for sandy soils, appearing in publications from
strains. They used the G/F(e)–p9 relation to obtain strain- the 1970s onwards, aimed at identifying the maximum shear
dependent parameters A and m, which generally exhibited modulus and the rate of shear modulus reduction with strain.
the same trend: A tends to decrease and m tends to increase Researchers have used both reconstituted specimens and
with strain. Index m is generally 0.5 from very small to high-quality undisturbed samples with equipment adapted to
small strains and then increases towards 1 for large strains. measure small-to-medium strains with good resolution.
Jovicic & Coop (1997) showed that both A and m have In order to collect all available shear modulus degradation
S-shaped relations with the logarithm of strain. curves and data from static and dynamic tests, a wide trawl
A number of researchers have advocated that the grain was done through the literature. More than 70 references
size and uniformity of sands affects their stiffness. Iwasaki were used, and more than 500 curves were digitised to create
& Tatsuoka (1977) defined, as an additional multiplier in a database. Eventually 454 tests were selected from 65
equation (1), a parameter B representing the influence of references. The relationship between normalised shear mod-
uniformity coefficient Uc , and fines content. Menq (2003) ulus (G/G0 ) and shear strain (ª) for the 454 selected tests is
proposed an additional multiplier depending on Uc , and plotted in Fig. 1. The corresponding 65 references with
suggested that index m was also affected. Hardin & Kalinski selected information from the database are summarised in
(2005) introduced a diameter parameter f (D) to calculate the Table 1. As seen in Fig. 1, points from each published test
maximum shear modulus for gravelly soils, but the dimen- are merged to create a curved but well-defined degradation
sional nature of this particular factor is problematic. zone. Three different procedures were used to generate these
The non-linear stress–strain behaviour of soils at small to points.
medium strains is mostly represented by some form of
(a) If a G–ª or G/G0 –ª curve was drawn using a solid line,
hyperbolic stress–strain relationship. Hardin & Drnevich
three points were located within each tenfold strain
(1972b) proposed this relationship as
interval starting from 106 strain.
G 1 (b) If G–ª or G/G0 –ª data comprised points with a number
¼ (2)
G0 1 þ ª=ªr of less than 20, then all the points were selected.
(c) If the number of points was more than 20, a similar
process was adopted as for solid lines.
where G is the secant shear modulus at any strain, G0 or
Gmax is the elastic (maximum) shear modulus (e.g. G at The scatter of points seen outside the general degradation
ª ¼ 0.0001%), and ªr is the reference shear strain, which is zone in Fig. 1 is very limited, and mainly in the very small
defined by max /G0 : The disadvantage of this approach is the strain range; these are usually single data points arising,
difficulty in finding max : The authors also indicated that this presumably, from inadequate resolution of strain. However,
true hyperbolic relationship did not generally fit their data. 5% of the published tests were excluded entirely. For
They therefore used a distorted strain scale, which they example, the paper of Iwasaki et al. (1978) included 36
called ‘hyperbolic shear strain’. tests, but only 35 of them were used; also Kokusho (1980)
Fahey & Carter (1993) reorganised the hyperbolic model published 18 tests, but two of them were excluded. The
so that modulus reduction became a function of shear excluded tests deviated strongly from the general trend. In
strength mobilisation, as seen in equation (3). In addition to some references, the excluded soils were thought to be
requiring max , this also uses empirical parameters f and g. cemented, but in others there was no obvious reason for the
They showed some success fitting this three-parameter model deviation. Leaking membranes or some other major but
to the data of a wide range of soils. unrecognised experimental problem may have occurred. As
can be seen from Fig. 2, the boundaries of the stiffness
G g
¼1 f (3) degradation zone, for those 95% of tests that have been
G0 max included, are very clearly defined.
The database covers a wide variety of sandy soils, includ-
Darendeli (2001) proposed a modified hyperbolic model ing dry, wet, saturated, reconstituted and undisturbed samples
based on testing of intact sand-gravel samples of clean sands, gravels, sands with fines and/or gravels, and
gravels with sands and fines, representing 60 different mater-
G 1
¼ a (4) ials (e.g. Toyoura sand, Ottawa sand, undisturbed Ishikari
G0 1 þ ª=ªr sand). The samples were prepared to investigate the effects
of changing the void ratio, relative density, anisotropy,
where a is called the curvature parameter, and ªr is the drainage conditions and confining pressure, and were tested
reference strain value at which G/G0 ¼ 0.50. This model under drained and/or undrained conditions in static and
uses only two parameters, and the reference strain provides dynamic tests. The various parameters and soil state/test
an efficient normalisation of shear strain. numbers are given for all references in Table 1. It is
In order to comprehend the non-linear elastic behaviour of suggested that these two columns should be evaluated
sands, and to produce a best-fit functional relationship, a together; the number in parentheses in the last column repre-
56 OZTOPRAK AND BOLTON
454 tests 124 tests
1·0 on dry-wet reconstituted 1·0 on dry samples
Normalised shear modulus, G/G0
0·4 0·4
0·2 0·2
0 0
0·0001 0·001 0·01 0·1 1 10 0·0001 0·001 0·01 0·1 1 10
Shear strain, γ: % Shear strain, γ: %
(a) (b)
0·6 0·6
0·4 0·4
0·2 0·2
0 0
0·0001 0·001 0·01 0·1 1 10 0·0001 0·001 0·01 0·1 1 10
Shear strain, γ: % Shear strain, γ: %
(c) (d)
sents the number of sand states, and it is followed by the ducted drained or undrained. This tends to confirm that
number of tests. For example, Chung et al. (1984) applied volume change effects have been negligible in their influence
three different confining pressures on the same sample, so it on shear stiffness.
is given as (1)3. In another row, Lo Presti et al. (1997) used Other influences are also too subtle to make it worth
six samples with different void ratios and tested them under extracting them. There are about ten samples whose OCR
the same confining pressure, so it is given as (6)6. value is between 1 and 5. Regarding these data, the effect of
Since it includes so many references and test numbers, it OCR on the degradation curve is apparently very limited.
is impractical to list all the details of the database. Fig. 3 Yamashita & Toki (1994) indicated that the shear modulus
was therefore prepared to illustrate the scope of the data- of sands at small strain levels is not affected by OCR,
base. The materials are mostly sands of various gradings, although a slight effect has been observed for medium
but mainly quite uniform, with some gravels; the relative strains. And, as Yamashita & Suzuki (1999) showed, whereas
density is mostly high but with some looser samples; and the direction of principal stress influences the elastic mod-
the confining pressures are mostly between 50 and 600 kPa, ulus G0 , it does not appear to change the G/G0 –ª curve.
with a median of 150 kPa. Caution must be exercised in Hence the normalised curve, given in Fig. 2, also allows for
using correlations based on the whole database to make the effects of anisotropy.
predictions for other materials that have ‘unusual’ sets of
parameter values that are sparsely represented in the data-
base. MODELLING SHEAR MODULUS DEGRADATION
Dynamic test data are more common in the literature, and The best-fit functional relationship for the secant shear
unavoidable for very small strains. The database accordingly modulus degradation data of Fig. 1 is shown in Fig. 2 as a
includes a considerable number of cyclic tests. However, an modified hyperbolic equation in the form
attempt was made to limit the numbers of cycles in a test to ," #
about ten, up to which any effects are generally found to be G ª ªe a
¼1 1þ (5)
negligible (Alarcon-Guzman et al., 1989; Yasuda & Matsu- G0 ªr
moto, 1994). Larger numbers of cycles may affect the shear
modulus degradation rate at medium to large strain levels, noting that for ª , ªe , G/G0 ¼ 1.0.
presumably because of the tendency of granular materials to
compact during cyclic loading. This same tendency causes (a) Lower bound: ªe ¼ 0; ªr ¼ 0.02%; a ¼ 0.88.
excess pore pressures to develop in undrained tests, which (b) Mean: ªe ¼ 0.0007% ; ªr ¼ 0.044%; a ¼ 0.88.
can ultimately lead to liquefaction. It is interesting to ob- (c) Upper bound: ªe ¼ 0.003%; ªr ¼ 0.10%; a ¼ 0.88.
serve, therefore, that Fig. 1 demonstrates no obvious ten- where G is the secant shear modulus at any strain; G0 is the
dency for bias between tests on dry or damp sand and those elastic (maximum) shear modulus (G at ª ¼ 0.0001%); ªr is
on saturated sand, irrespective of whether they were con- the characteristic reference shear strain (shear strain at
Table 1. Summary of the database used for this study
Reference Soil materials and description Test type and drainage Changed factors Number of tests
Alarcon-Guzman et al. (1989) Ottawa 20/30 sand (dry, vibrated) RCT+TST, RCT e, OCR, stress ratio, p9 (8)11
Arango (2006) Calcareous and silica sands (dry, remoulded) RCT e, p9 (5)17
Cavallaro et al. (2003) Noto soil (undisturbed lightly cemented silty sand) RCT+CLTST (U) e, p9 (2)2
Charif (1991) Hostun sand (compacted) TXT (D) e, p9 (3)5
Chegini & Trenter (1996) Glacial till (undisturbed, includes coarse material) RCT (U) PI, stone content, p9 (4)7
Chung et al. (1984) Monterey sand No. 0 CLTXT (U) p9 (1)3
Delfosse-Ribay et al. (2004) Fontainebleau sand (with/without sodium silicate grouted) RCT+CLTXT (U) p9 (2)2
Dong et al. (1994) Hime gravel (air-dried, compacted) MLTXT (D) e, D50 , Uc (2)2
D’Onofrio & Penna (2003) Silty sand (compacted) RCT, CLTST (D) e, p9 (4)7
Drnevich & Richart (1970) Ottawa 30/50 sand (dry) RCT e, p9 (1)3
Ellis et al. (2000) Fine and coarse silica sand (dry), Toyoura sand (dry and saturated) RCT e, D50 , Uc (2)2
Fioravante et al. (1994) Quiou sand (carbonatic, low-density reconstituted) RCT (D) e, OCR, p9 (5)7
Goto et al. (1987) Gravel (undisturbed and reconstituted) CLTXT (U) ID , p9 (2)2
Goto et al. (1992) Gravel (undisturbed and reconstituted, 24% and 44% sand content) CLTXT (U) Sampling depth, Go (4)4
Hardin & Drnevich (1972a) Sand (clean and dry) CLSST p9 (5)5
Hardin & Drnevich (1972b) Sand (clean and dry) CLSST p9 (1)2
Hardin & Kalinski (2005) Ottawa sand, sands (dry), gravel–sand–silt mixture (wet) RCT (D) e, p9 (2)6
Hashiba (1971) Onahama sand (dry, remoulded) SST e, p9 (2)4
Hatanaka & Uchida (1995) Tokyo gravel (undisturbed and reconstituted, sandy, includes fines) CLTXT (U) Go (2)2
Hatanaka et al. (1988) Tokyo gravel (undisturbed and reconstituted, sandy, includes fines) CLTXT (U) p9 (2)4
Ishihara (1996) Fujisawa sand (undisturbed and reconstituted) TST (U) e (4)4
Ito et al. (1999) Sandy gravel (undisturbed and improved by grouting) CLTXT (D) e, D50 , p9 (4)4
Iwasaki et al. (1978) Toyoura, Ban-nosu, Iruma, Kinjo-1, Kinjo-2, Ohgi-Shima, Monterey sands RCT, TST (D) e, p9 (23)35
Jardine et al. (2001) Dunqerque sand TXT, TST (D) OCR, test type (2)3
Katayama et al. (1986) Sand (undisturbed and reconstituted) RCT Sampling depth, Go (4)4
Koga et al. (1994) Gravelly soil (dry, reconstituted) CLTST Gravel content (5)5
Kokusho (1980) Gifu and Toyoura sands (compacted, saturated) CLTXT Drainage, e, p9 (10)16
Kokusho & Esashi (1981) Toyoura sand (saturated), undisturbed diluvial sand, gravel (saturated) CLTXT (U) e, p9 (3)11
Kokusho & Tanaka (1994) Gravel (undisturbed and reconstituted) CLTXT (U) p9 (7)7
Laird & Stokoe (1993) Sand (dry, remoulded) CLTXT, RCT (U) p9 (1)6
Lanzo et al. (1997) Santa Monica and Antelope Valley sands (reconstituted, includes silt) DSST (D) e, OCR, p9 (6)10
Lo Presti et al. (1997) Toyoura and Quiou sands (dry, hollow cylinder) MLTST e (6)6
Maher et al. (1994) Ottawa sand (medium and loose, untreated and chemically grouted) RCT + CLTXT ID , concentration, curing time (8)11
Menq (2003) Washed mortar sand, sand and gravelly sand (dry) RCT e, D50 , Uc (7)7
Menq & Stokoe (2003) Sand (dense and dry) RCT p9 (1)3
Ogata & Yasuda (1982) Gravelly soil (undisturbed, includes fines) TXT p9 (1)2
Park (1993) Toyoura sand (air-dried, isotropic consolidated) PSCT, TXT OCR, p9 (2)2
STIFFNESS OF SANDS THROUGH A LABORATORY TEST DATABASE
Porovic & Jardine (1994) Ham River sand (K0 and isotropic consolidated) RCT+TST (U) Consolidation pressure, e, OCR (8)8
Rollins et al. (1998) Sand and gravelly sand (compacted, saturated) RCT (U) e, gravel content, p9 (1)5
Saxena et al. (1988) Monterey sand (compacted, saturated, with/without cement, curing) RCT (D) ID , cement content, curing time, p9 (14)22
Seed et al. (1986) Gravel (Oroville and Pyramid materials) CLTXT (U) e (2)2
Shibuya et al. (1996) Higashi-Ohgijima sand (undisturbed) CLTXT, TST (U) Sampling depth, p9 (8)8
Silver & Seed (1971) Quartz sand (crystal silica No. 20, dry) CLSST p9 (1)3
Stokoe et al. (1994) Sand (undisturbed and dry-remoulded) CLTST, RCT e, p9 (5)9
Stokoe et al. (1999) Silty sand (undisturbed) RCT Sampling depth, p9 (4)4
Tanaka et al. (1987) Gravelly sand CLTXT Gravel content, p9 (2)6
Teachavorasinskun et al. (1991) Toyoura sand (dense) TSST (D) e, OCR (2)2
57
58 OZTOPRAK AND BOLTON
1·0
G 1
⫽
(4)16
(9)15
(8)14
⎝ a
⎛ γ ⫺ γe
(6)6
(2)2
(2)2
(2)2
(9)9
(2)4
(1)4
(3)8
G0 ⎜
0·8 1⫹⎜
γr ⎛
⎝
0·6
Upper bound
Consolidation condition, drainage
0·4
Lower bound
Consolidation pressure, e, w
0·2
CL, cyclic loading; ML, monotonic loading; RCT, resonant column test; TXT, triaxial test; TST, torsional shear test; SST, simple shear test; TSST, torsional simple shear test.
0
0·0001 0·001 0·01 0·1 1 10
Drainage, Go , p9
Shear strain, γ: %
e, D50 , Uc , p9
e, D50 , Uc , p9
e, D50 , Uc , p9
(a)
e, drainage
D50 , Uc
Lower Upper
Go , p9
e, p9
p9
e
Mean
0·6
CLTSST (D)
CLTXT (U)
CLTXT (U)
CLTXT (D)
CLTXT (D)
Upper bound
RCT (U)
RCT (D)
TXT (U)
RCT (U)
0·4
CLTXT
RCT
RCT
Lower bound
0·2
0
0·0001 0·001 0·01 0·1 1 10
Shear strain, γ: %
(b)
Gravel (undisturbed and reconstituted, loading and unloading during test)
Toyoura sand (remoulded, isotropically or anisotropically consolidated)
m(ª)
AðªÞ pa p9
G¼ 3
(6)
ð1 þ eÞ pa
80 80
Test number
Test number
120 60 60
80 40 40
40 20 20
0 0 0
0·25–0·42
⬍50
50–100
100–200
200–400
400–600
⬎600
0·25–0·40
0·65–0·70
0·70–0·80
0·42–0·60
0·60–0·80
0·80–0·90
0·10–0·25
0·40–0·55
0·55–0·65
0·80–0·95
0·95–1·15
⬍0·25
0·90–0·95
⬎0·95
Confinement pressure: kPa Void ratio, e Relative density, ID
(a) (b) (c)
120 363 tests 416 tests 300 For 328 wet and
100
100 Average ⫽ 4·2 mm Average ⫽ 8·4 250 UD samples
Test number
Test number
Test number
80 200
60
60 150
40 40 100
20 20 50
0 0 0
Drained
Undrained
UD
0·10–0·14
0·14–0·20
0·20–0·30
0·30–0·70
0·70–1·50
1·50–10·0
10·0–20·0
⬎20·0
1·95–5·0
5·0–10·0
10·0–30·0
⬎30·0
⬍1·15
1·15–1·35
1·35–1·55
1·55–1·95
Dry
Wet
NA
Average grain size, D50: mm Uniformity coefficient, Uc Sampling
(d) (e) (f)
p9. In order to investigate this, equations (5) and (6), which modify the A(ª) and m(ª) values slightly from the values
are alternative methods of representing shear modulus calcu- listed in Fig. 4 so as to fit exactly the mean curve in Fig.
lations at any strain level, can be compared. 5(a). The refined values appear in Fig. 6.
To plot the results of equation (6) on the mean curve of Once the calculated G/G0 values of equation (5) have
G/G0 against ª, p9 in equation (6) can be assigned as been fitted to the mean hyperbolic curve in Fig. 5(a) for the
150 kPa, which is the average value of the median stress mean pressure p9 ¼ 150 kPa, other p9 values can be inserted
range in Fig. 3. Since G/G0 is to be calculated, the void in equation (6), and appropriate values of ªe and ªr can be
ratio function cancels, and is not required. Using p9, A(ª) derived from equation (5) to fit the new curves. The result-
and m(ª) from Fig. 4, the G/G0 values can readily be ing family of degradation curves for varying p9 is shown in
calculated and placed on Fig. 5(a), where they lie very close Fig. 5(b). Fig. 6(b), accumulating the results of 379 tests,
:
to the mean curve, but not exactly on it. Since the normal- shows that G0 can best be expressed as a function of (p9)0 5
ised shear modulus curve gives a more reliable relation than at very small strains and as a function of p9 for large strains.
the scattered relationship shown in Fig. 4, it is preferred to If the resulting m(ª) relation is compared with the published
100 000
10 000
GV 3/pa
1000
100 000
γ: % A(γ) m(γ)
10 000
0·0001 5760 0·49
GV 3/pa
Fig. 4. Shear modulus variation with mean effective stress: (a) ª 0.0001% (379 tests); (b) ª 0.001% (379 tests); (c) ª 0.01% (374
tests); (d) ª 0.1% (221 tests); (e) ª 1% (77 tests)
60 OZTOPRAK AND BOLTON
1·0
A(γ)
0·4 Eq. (6) G ⫽ (pa)1⫹m(γ) p⬘m(γ)
(1 ⫹ e)3
0
0·0001 0·001 0·01 0·1 1 10
γ: %
(a)
1·0
0·6
G /G0
p⬘: MPa
0·05
0·4
0·15
0·4
0·2 0·8
Points calculated using eq. (6).
2 A and m values are modified values
4 given in Fig. 6. p⬘ ⫽ 150 kPa
0
0·0001 0·001 0·01 0·1 1 10
γ: %
(b)
Fig. 5. Calibration of m(ª) and A(ª) parameters using equation (6) (symbols) with equation
(5) (curves)
data of other researchers, as shown in Fig. 7, it now creates (70 kPa to 600 kPa), the sands typical of this database
a trend line appropriate to most sands, verifying the mod- produce regressions for ªe and ªr as follows.
ification process explained above.
p9
ªe (%) ¼ ð8 3 105 Þ þ 6 3 104 (7)
pa
Reference strain and elastic threshold strain p9
According to the database curve in Fig. 2, G/G0 is 0.5 at ªr (%) ¼ 0:008 þ 0:032 (8)
pa
a shear strain between 0.02% and 0.1%, with a mean of
0.044%. Using a constant value a ¼ 0.88 in equation (5), Darendeli (2001) and Menq (2003) suggested a power
and changing the confining pressure progressively from relation between ªr and p9. However, they both indicated
10 kPa to 4 MPa, leads to the interesting offsetting of the that this relation tends to become linear for higher stresses.
modulus degradation curves towards higher values of strain, Menq (2003) also introduced the effect of uniformity coeffi-
as seen in Fig. 5(b). The trend of ªe and ªr increasing with cient for sands and gravels, and proposed an equation as
confining stress is shown in normalised form in Fig. 8. Both follows.
ªe and ªr exhibit a logarithmic function with stress in the 0:5U 0 :15
:6 p9
c
low-stress range (p9 , 70 kPa). In the intermediate stress ªr ¼ 0:12U c0
(9)
range (70 kPa , p9 , 600 kPa) the relation is nearly linear, pa
but the rate of increase reduces slightly at even higher
stresses. It must be presumed that grain contact deformation Unlike the approaches of Darendeli (2001) and Menq
and damage are responsible for increasing characteristic (2003), it is considered convenient here to define the stress
strains in this way. For the most common confining stresses level as low, medium and high, and to accept a linear
STIFFNESS OF SANDS THROUGH A LABORATORY TEST DATABASE 61
6000
respective trials, and possibly the influence of water at
the smallest confining stresses. However, the gradients for
5000
γ: % A(γ) this material in Fig. 10 are identical.
4000 0·0001 5760
To generalise the relation between ªr and p9, the database
0·001 5520 was searched for all those soils on which tests were con-
ducted at three or more confining pressures, providing data
A(γ)
0·003 5230
3000
0·01 4520 for the evolution of ªr : This identified 24 sandy soils, and
0·03 3150
2000 0·1 1810 their data were split into four groups, as shown in Fig. 11. It
0·3 880 can be confirmed that almost all the plots of ªr exhibit a
1 370 linear trend with increasing mean effective stress. As seen
1000 3 180
10 126 from Fig. 11, both uniformity and relative density exert a
0 significant influence on the ªr –p9/pa relation. Fig. 11 there-
0·0001 0·001 0·01 0·1 1 10 fore represents the functional dependence of ªr on p9, ID
γ: %
and Uc for different states and types of sandy soils.
(a)
Various regressions are given in Fig. 12. Fig. 12(a) shows
1·1
γ: % m(γ)
a power relation between ªr and void ratio e, albeit with a
1·0
moderate coefficient of determination R2 ¼ 0.54. Fig. 12(b)
0·0001 0·50
0·001 0·51 shows negligible correlation between ªr and relative density
0·9 0·003 0·53 ID , but Fig. 12(c) demonstrates a much-improved R2 ¼ 0.74
0·01 0·56 for a power law correlation between ªr and the product eID :
0·8 0·03 0·63
This was initially unexpected. However, it is proposed that
m(γ)
0·1 0·73
0·7 0·3 0·83 the group eID may be a surrogate for grain shape, which is
1 0·93 the most significant omission from the current study of sand
3 0·99
0·6
10 1·00
characteristics in relation to stiffness degradation. High void
ratio for a rounded sand would indicate low relative density,
0·5
so the product eID would also be small, whereas an angular
0·4 sand could have a high void ratio at a high relative density
0·0001 0·001 0·01 0·1 1 10 and give a large product eID : So the magnitude of eID may
γ: % indicate angularity. This possible explanation cannot be
(b) verified, since the authors whose work we have used did not
generally remark on grain shape. Nevertheless, the statistical
Fig. 6. Modified relations of m(ª) and A(ª)
finding is significant.
If one equation will cover all these effects for calculating
ªr , it must be expressed in the form
relation for medium and high stress levels. Moreover, a
p9
linear relation offers a better fit to the data than a power ªr (%) ¼ c þd (10)
relation, at least for medium stress levels (70–600 kPa). pa
Further support for a linear relation between ªr and p9
comes from the normalised shear modulus degradation of for medium stress levels. From Fig. 11, it is understood that
Toyoura sand reported by Iwasaki et al. (1978) and Kokusho Uc affects the slope of the ªr –p9/pa relation, and from Figs
(1980), and plotted in Fig. 9. The corresponding changes in 10, 11 and 12 that the product eID affects the ordinate.
ªr are plotted in Fig. 10 for the stress range 50–300 kPa; Accepting these functional relationships, a multivariable
the trend is obviously linear. The presumed cause of the regression analysis for medium stress levels then produced
different intercepts in Fig. 10 is the relative density of the the relation
1·1
1·0
Yasuda & Matsumoto (1993), gravel, undr. test,
e ⫽ 0·42, ID ⫽ 0·70, D50 ⫽ 14 mm, Uc ⫽ 7
0·9
This study
Iwasaki et al. (1993), Toyoura sand, saturated,
0·8 dr test, e ⫽ 0·68, ID ⫽ 0·88, D50 ⫽ 0·16 mm, Uc ⫽ 1·46
Silver & Seed (1971)
silica sand No. 20, dry,
m(γ)
0·03 0·0006
γr: %
γe: %
0·02 0·0004
0·01 0·0002
10 kPa
70 kPa
10 kPa
70 kPa
y ⫽ 0·009ln(x) ⫹ 0·037 y ⫽ 8 ⫻ 10⫺5ln(x) ⫹ 0·0007
R 2 ⫽ 0·9977 R 2 ⫽ 0·9983
0 0
0 0·2 0·4 0·6 0·8 0 0·2 0·4 0·6 0·8
p' /pa p' /pa
(a) (b)
0·10 0·0015
0·08 0·0012
0·06 0·0009
γr: %
γe: %
0·04 0·0006
600 kPa
600 kPa
70 kPa
70 kPa
0·02 y ⫽ 0·008x ⫹ 0·032 0·0003 y ⫽ 8 ⫻ 10⫺5x ⫹ 0·0006
R 2 ⫽ 0·9983 R 2 ⫽ 0·9991
0 0
0 2 4 6 0 2 4 6
p' /pa p' /pa
(c) (d)
0·4 0·005
0·004
600 kPa
0·3
600 kPa
0·003
γe: %
γr: %
0·2
0·002
0·1
4 MPa
4 MPa
Fig. 8. Trend for evolution of ªe and ªr for different confining pressure ranges, derived from Fig. 5(b)
0:3 p9
ªr (%) ¼ 0 01U c
: þ 0:08eI D (11) likely to involve elastic contact deformation mediated by a
pa
greater amount of grain rotation, and a reduced proportion
It is interesting to reflect on the physical origins of of contact sliding. It may therefore be concluded that equa-
granular behaviour that could lead to the parametric influ- tion (11) is in accordance with a micromechanical under-
ences in equation (11). An increase in uniformity coefficient standing of soil behaviour.
Uc leads to a reduction in ªr – that is, to a swifter loss of Elastic threshold strain marks the onset of non-linearity,
elastic stiffness with strain. McDowell & Bolton (1999) and is therefore associated with the onset of contact sliding.
discussed the consequences of a dispersion of particle sizes Identical micromechanical considerations apply to ªe as to
in terms of the strain incompatibility between the fine matrix ªr ; it must therefore be anticipated that they will be corre-
and the larger particle. The introduction of large particles lated. The value of ªe can be extracted from the database
inevitably causes premature sliding of smaller particles in using the best-fit modulus reduction curve as expressed by
contact with them. Since a sliding contact no longer con- equation (5). Furthermore, it is clear in Fig. 5(b) that ªe
tributes its tangential shear stiffness to the global shear increases at G/G0 ¼ 1 during an increase in the mean effec-
modulus, the onset of sliding coincides with the reduction of tive stress, and that it has the same trend as ªr , albeit at a
G/G0 : In other words, the reference strain ªr is reduced much smaller strain magnitude: this was set out in Fig. 8.
when there is a greater disparity in grain sizes. Equation Fig. 13 demonstrates that a simple linear relation can be
(11) supports this finding. derived between ªe and ªr (expressed in percentages) using
According to equation (11), it must also be accepted that all the available data, as
increased mean effective stress p9 or increased relative ªe ¼ 0:0002 þ 0:012ªr (12)
density ID (noting that an increase in ID overwhelms the
concomitant reduction in e) tends to protect the granular
material somewhat from the reduction of stiffness due to
strain: ªr increases. This might be attributed in both cases to Curvature parameter
increased interlocking. Increased p9 will lead to increased The curvature parameter a varies from 0.75 to 1.0 in the
contact flattening, and a tendency to suppress degrees of database, and 0.88 is the average value for uncemented
freedom associated with sliding. Increasing ID also wedges sands, as employed in equation (5). Darendeli (2001) sug-
more grains in place. In these cases, small strains are more gested a constant value of 0.92 for a smaller range of
STIFFNESS OF SANDS THROUGH A LABORATORY TEST DATABASE 63
1·0 0·25
0·20
0·8
0·15
γr: %
γr
0·6
G /G 0
0·10
0·4
200 kPa
0·05 Mean: (R 2 ⫽ 0·7636)
Toyoura sand (Iwasaki et al., 1978) y ⫽ 0·037x ⫹ 0·030
0·2 Dry, resonant column test 0
100 0 1 2 3 4
e ⫽ 0·68, ID ⫽ 0·80, D50 ⫽ 0·16 mm, Uc ⫽ 1·46
25 50 p⬘/pa
0
(a)
0·0001 0·001 0·01 0·1 1
0·25
γ: %
(a)
0·20
1·0
0·15
γr: %
0·8
0·10
γr
0·6 0·05 Mean: (R 2 ⫽ 0·8032)
G /G0
y ⫽ 0·020x ⫹ 0·027
0·4 300 kPa 0
0 1 2 3 4
200 p⬘/pa
0·2 Toyoura sand (Kokusho, 1980) (b)
100
Saturated, drained CLTXT 50 0·25
e ⫽ 0·64, ID ⫽ 0·92, D50 ⫽ 0·184 mm, Uc ⫽ 1·80 20
0
0·0001 0·001 0·01 0·1 1 0·20
γ: %
(b) 0·15
γr: %
Mean: (R 2 ⫽ 0·5466)
0·10 y ⫽ 0·012x ⫹ 0·022
Fig. 9. Normalised shear modulus degradation curves of Toyoura
sand shifting with mean effective stress
0·05
0·20 0
y ⫽ 0·032x ⫹ 0·0547 0 1 2 3 4
p⬘/pa
R2 ⫽ 0·9993
0·15 (c)
Kohusho (1980) 0·25
γr: %
0·20
0·10
y ⫽ 0·032x ⫹ 0·018
0·15
? R2 ⫽ 1·0
γr: %
0·05
Iwasaki et al. (1978) 0·10
Mean: (R 2 ⫽ 0·1302)
y ⫽ 0·002x ⫹ 0·015
0·05
0
0 1 2 3
p⬘/pa 0
0 1 2 3 4
Fig. 10. Evolution of ªr for Toyoura sand at increasing p9 p⬘/pa
(d)
0·003
γr: %
0·1
γe: %
0·002
0·001
0·01 y ⫽ 0·012x ⫹ 0·0002
0·1 1 10
e R 2 ⫽ 0·9978
(a) 0
0 0·1 0·2 0·3
1
y ⫽ 0·0562x0·3635 γr: %
R 2 ⫽ 0·0355
Fig. 13. Relation between ªr and ªe
10
γr: %
0·1
y ⫽ 1·0166x0·223
R 2 ⫽ 0·6952
a 1
0·01
0·1 1
ID
(b)
1
0·1
y ⫽ 0·1236x1·0675 0·1 1
R 2 ⫽ 0·7353 e
(a)
10
y ⫽ 0·8498x⫺0·0416
R 2 ⫽ 0·5141
γr: %
0·1
a 1
0·01
0·1 1
eID
(c) 0·1
1 0·1 1 10 100
y ⫽ 0·0714x⫺0·3293 D50
R 2 ⫽ 0·7376 (b)
10
y ⫽ 0·9767x⫺0·0746
R 2 ⫽ 0·8748
γr: %
0·1
a 1
0·01
1 10 100
Uc 0·1
(d) 1 10 100
Uc
Fig. 12. Influence on ªr of void ratio e, relative density ID and (c)
uniformity coefficient Uc
Fig. 14. Influence on curvature parameter a of void ratio and
uniformity coefficient
eter for the G/G0 –ª curves of the whole database of sandy
soils in Fig. 2, the statistical analysis of variations between vary very much, their influence on soil stress–strain curves
the characteristics of the soils and their test conditions has is by no means insignificant. Fig. 15(a) translates from G
resulted in the more refined expression in equation (13). against ª at mean effective stress p9 ¼ 100 kPa into shear
Although parameters such as ªr and a may not appear to stress against ª in a hypothetical simple shear test. The
STIFFNESS OF SANDS THROUGH A LABORATORY TEST DATABASE 65
1000
elastic stiffness at very small strains is taken as a constant
G0 ¼ 250 MPa. In Fig. 15(a) it is shown that, for a typical
sand with a ¼ 0.88, the influence of reference strain ªr in
the range 0.02–0.1% creates a fourfold variation in the
mobilisation of shear stress up to 1% shear strain. Fig. 15(b)
shows a more modest, but nevertheless significant, variation
G0 (calculated): MPa
in expected mobilised shear stress due to variations of a
within the typical range 0.80–1.0, for the average value of
ªr ¼ 0.044%. 100
VALIDATION
In the case of unavailability of the linear elastic shear
modulus value and its reduction by straining for a sandy Factor 2
soil, it is possible to calculate them with equations (5), (11),
(12) and (13) proposed in this paper. Comparisons between
measured and predicted values can be validated against the 10
database. In Fig. 16(a) it is shown that 86% of the 345 10 100 1000
calculated values of G0 lie within a factor of 2 of the G0 (measured): MPa
(a)
measured values, implying a standard deviation of a factor 1
of 1.6 if the variation is normally distributed. Fig. 16(b)
shows that 94% of the 194 calculated values of reference
strain ªr lie within a factor of 2 of the measured values,
implying a standard deviation of a factor of 1.4. And Fig.
16(c) shows that all 280 calculated values of curvature
parameter a effectively lie within a factor of 1.3 of the
γr (calculated): %
interpreted measurements.
The overall significance of the residual deviations in
0·1
modulus reduction, G/G0 , between predictions and measure-
ments can best be assessed by plotting predicted against
Factor 2
measured values for all 3860 data points accumulated from
all the tests, on those soils in the new database that are
sufficiently well classified to enable the comparison. This is
γr ⫽ 0·1%
300 48
Moblised friction angle, φ ⬘mob: degrees
0·01
250 0·01 0·1 1·0
γr (measured): %
Shear stress, τ: kPa
(b)
200
10
γr ⫽ 0·44%
150 37
100 γr ⫽ 0·02%
25
50
a (calculated)
Factor 1·3
0 1
0 0·2 0·4 0·6 0·8 1·0
Shear strain, γ: %
(a)
300
Moblised friction angle, φ ⬘mob: degrees
250 a ⫽ 0·8
44
Shear stress, τ: kPa
200 0·1
0·1 1 10
a ⫽ 0·9 a (measured)
150 36 (c)
a ⫽ 1·0
100 31
Fig. 16. Comparison of measured and calculated values of (a) G0
(345 tests), (b) ªr (194 tests) and (c) a (280 tests)
50
1·0 1·0
Test data Test data
Database curves Database curves
0·8 Predicted 0·8 Predicted
0·6 0·6
G/G0
G/G0
0·4 0·4
Fioravante et al. (1994) Yamashita & Toki (1995)
Saturated Quiou sand Undisturbed (UD) Ishkari sand
0·2 0·2
Drained RCT, p⬘ ⫽ 250 kPa Undrained CLTXT, p⬘ ⫽ 180 kPa
e ⫽ 1·04, ID ⫽ 0·53, Uc ⫽ 1·3 e ⫽ 0·928, ID ⫽ 0·84, Uc ⫽ 1·7
0 0
0·0001 0·001 0·01 0·1 1 10 0·0001 0·001 0·01 0·1 1 10
γ: % γ: %
(a) (b)
1·0 1·0
Test data Test data
Database curves Database curves
0·8 Predicted 0·8 Predicted
0·6 0·6
G/G0
G/G0
0·4 0·4 Alarcon-Guzman et al.
Yamashita & Toki (1994) (1989)
Saturated Toyoura sand Dry Ottawa sand
0·2 0·2
Undrained CLTXT, p⬘ ⫽ 98 kPa RCT ⫹ TST, p⬘ ⫽ 100 kPa
e ⫽ 0·688, ID ⫽ 0·80, Uc ⫽ 1·3 e ⫽ 0·635, ID ⫽ 0·40, Uc ⫽ 1·2
0 0
0·0001 0·001 0·01 0·1 1 10 0·0001 0·001 0·01 0·1 1 10
γ: % γ: %
(c) (d)
1·0 1·0
Test data Test data
Database curves Database curves
0·8 Predicted 0·8 Predicted
0·6 0·6
G/G0
G/G0
1·0 1·0
Test data Test data
Database curves Database curves
0·8 Predicted 0·8 Predicted
0·6
G/G0
0·6
G/G0
0·4 0·4
Ito et al. (1999) Dong et al. (1994)
Sandy gravel Hime gravel, compacted
0·2 0·2
Undrained CLTXT, p⬘ ⫽ 100 kPa Drained MLTXT, p⬘ ⫽ 78·5 kPa
e ⫽ 0·38, ID ⫽ 0·85, Uc ⫽ 30 e ⫽ 0·73, ID ⫽ 0·72, Uc ⫽ 1·46
0 0
0·0001 0·001 0·01 0·1 1 10 0·0001 0·001 0·01 0·1 1 10
γ: % γ: %
(g) (h)
1·0 1·0
Test data Test data
Database curves Database curves
0·8 Predicted 0·8 Predicted
0·6 0·6
G/G0
G/G0
0·4 0·4
Rollins et al. (1998) Yasuda et al. (1996)
Saturated sand (GC ⫽ 20%) Undisturbed (UD) gravel
0·2 0·2
Undrained RCT, p⬘ ⫽ 200 kPa Drained CLTXT, p⬘ ⫽ 590 kPa
e ⫽ 0·685, ID ⫽ 0·40, Uc ⫽ 36 e ⫽ 0·25, ID ⫽ 0·70, Uc ⫽ 70
0 0
0·0001 0·001 0·01 0·1 1 10 0·0001 0·001 0·01 0·1 1 10
γ: % γ: %
(i) (j)
Fig. 18. Verification of predictions using equations (5), (11), (12) and (13) against shear modulus reduction curves of
various sandy soils (UD, undisturbed; GC, gravel content)
68 OZTOPRAK AND BOLTON
a factor of 1.13 for one standard deviation of random error, soil. In Deformation characteristics of geomaterials (eds H. Di
as determined from 3860 data points. This very narrow Benedetto, T. Doanh, H. Geoffroy and C. Saizéat), pp. 267–274.
spread applies irrespective of the test type and soil condi- Lisse, the Netherlands: Swets & Zeitlinger.
tion, within the range listed in Table 1. The initial elastic Charif, K. (1991). Contribution à l’étude du comportement mécani-
shear modulus, G0 , should always be measured if possible, que en petites et grandes déformation ( . 10–6 ). PhD disserta-
tion, Ecole Centrale Paris, France (in French).
but a new empirical relation is shown to provide estimates Chegini, A. & Trenter, N. A. (1996). The shear strength and
within a factor of 1.6 for one standard deviation of random deformation behaviour of a glacial till. In Advances in site
error, as determined from 379 tests. investigation practice (ed. C. Craig), pp. 851–866. London, UK:
Thomas Telford.
Chung, R. M., Yokel, F. Y. & Drnevich, V. P. (1984). Evaluation of
ACKNOWLEDGEMENTS dynamic properties of sands by resonant column testing. Geo-
The work was supported by EPSRC Platform Grant tech. Test. J. 7, No. 2, 60–69.
GR/T18660/01. The first author was also supported by the Clayton, C. R. I. (2011). Stiffness at small strain: research and
Scientific Research Projects Coordination Unit of Istanbul practice. 50th Rankine Lecture. Géotechnique 61, No. 1, 5–37,
http://dx.doi.org/10.1680/geot.2011.61.1.5.
University, Project No. YADOP-16527.
Darendeli, B. M. (2001). Development of a new family of normal-
ized modulus reduction and material damping curves. PhD
dissertation, University of Texas at Austin, TX, USA.
NOTATION Delfosse-Ribay, E., Djeran-Maigre, I., Cabrillac, R. & Gouvenot, D.
A material constant (2004). Shear modulus and damping ratio of grouted sand. Soil
a curvature parameter Dynam. Earthquake Engng 24, No. 6, 461–471.
B material constant Dong, J., Nakamura, K., Tatsuoka, F. & Kohata, Y. (1994). Defor-
c scaling coefficient mation characteristics of gravels in triaxial compression tests
D diameter and cyclic triaxial tests. In Pre-failure deformation of geomater-
D50 average grain size ials (eds S. Shibuya, T. Mitachi and S. Miura), pp. 17–23.
d scaling coefficient Rotterdam, the Netherlands: Balkema.
e void ratio D’Onofrio, A. & Penna, A. (2003). Small strain behaviour of a
eg empirical parameter in equation (1) lime-treated silty sand. In Deformation characteristics of geo-
emax maximum void ratio materials (eds H. Di Benedetto, T. Doanh, H. Geoffroy and C.
emin minimum void ratio Saizéat), pp. 329–336. Lisse, the Netherlands: Swets & Zeit-
f empirical parameter linger.
G secant shear modulus; shear stiffness Drnevich, V. P. & Richart, F. E. (1970). Dynamic prestraining of
Gmax maximum shear modulus dry sand. J. Soil Mech. Found. Div. ASCE 96, No. SM2, 453–
G0 initial elastic shear modulus 469.
g empirical parameter Ellis, E. A., Soga, K., Bransby, M. F. & Sato, M. (2000). Resonant
ID relative density column testing of sands with different viscosity pore fluids.
m material constant J. Geotech. Geoenviron. Engng 126, No. 1, 10–17.
N60 corrected blow count Fahey, M. & Carter, J. P. (1993). A finite element study of the
OCR overconsolidation ratio pressuremeter test in sand using a nonlinear elastic plastic
PI plasticity index model. Can. Geotech. J. 30, No. 2, 348–362.
p9 mean effective stress Fioravante, V., Jamiolkowski, M. & Lo Presti, D. C. F. (1994).
pa reference (atmospheric) pressure Stiffness of carbonatic Quiou sand. Proc. 8th Int. Conf. Soil
pc preconsolidation pressure Mech. Found. Engng, New Delhi 1, 163–168.
R2 coefficient of determination Gibbs, H. J. & Holtz, W. G. (1979). Discussion of ‘SPT and relative
s9 effective stress density in coarse sands’, by Marcuson and Bieganousky.
Uc uniformity coefficient J. Geotech. Engng ASCE 105, No. GT3, 439–441.
V ¼ 1+e Goddard, J. D. (1990). Nonlinear elasticity and pressure-dependent
w water content wave speeds in granular media. Proc. R. Soc. Math. Phys. Sci.
ª shear strain 430, No. 1878, 105–131.
ªe elastic threshold strain Goto, S., Syamoto, Y. & Tamaoki, K. (1987). Dynamic properties
ªr reference strain of undisturbed gravel samples obtained by the in-situ freezing
v90 vertical effective stress method. Proc. 8th Asian Regional Conf. on Soil Mech. Found.
19 major effective principal stress Engng, Kyoto 1, 233–236.
29 intermediate effective principal stress Goto, S., Suzuki, Y., Nishio, S. & Hiroshi, O.-O. (1992). Mechani-
39 minor effective principal stress cal properties of undisturbed Tone-River gravel obtained by in-
shear stress situ freezing method. Soils Found. 32, No. 3, 15–25.
max maximum shear stress Hardin, B. O. & Black, W. L. (1966). Sand stiffness under various
9mob mobilised friction angle triaxial stresses. J. Soil Mech. Found. Div. ASCE 92, No. SM2,
667–692.
REFERENCES Hardin, B. O. & Drnevich, V. P. (1972a). Shear modulus and
Alarcon-Guzman, A., Chameau, J. L., Leonards, G. A. & Frost, J. damping in soils: measurement and parameter effects. J. Soil
D. (1989). Shear modulus and cyclic undrained behavior of Mech. Found. Div. ASCE 98, No. SM6, 603–624.
sands. Soils Found. 29, No. 4, 105–119. Hardin, B. O. & Drnevich, V. P. (1972b). Shear modulus and
Arango, J. C. (2006). Stress–strain behavior and dynamic properties damping in soils: design equations and curves. J. Geotech.
of Cabo Rojo calcareous sands. MSc thesis, University of Engng 98, No. 7, 667–692.
Puerto Rico, San Juan, Puerto Rico. Hardin, B. O. & Kalinski, M. E. (2005). Estimating the shear
Atkinson, J. H. (2000). Nonlinear soil stiffness in routine design. modulus of gravelly soils. J. Geotech. Geoenviron. Engng 131,
Géotechnique 50, No. 5, 487–508, http://dx.doi.org/10.1680/ No. 7, 867–875.
geot.2000.50.5.487. Hashiba, T. (1971). Simple shear apparatus using an inclinometer.
Atkinson, J. H. & Salfors, G. (1991). Experimental determination of Soils Found. 11, No. 3, 113–119.
stress–strain–time characteristics in laboratory and in situ tests. Hatanaka, M. & Uchida, A. (1995). Effect of test methods on the
Proc. 10th Eur. Conf. Soil Mech. Found. Engng, Florence 3, cyclic deformation characteristics of high quality undisturbed
915–956. gravel samples. In Static and dynamic properties of gravelly
Cavallaro, A., Maugeri, M. & Ragusa, A. (2003). Small strain soils (eds M. D. Evans and R. J. Fragaszy), Geotechnical Special
stiffness from in situ and laboratory tests for the city of Noto Publication No. 56, pp. 136–161. Reston, VA, USA: ASCE.
STIFFNESS OF SANDS THROUGH A LABORATORY TEST DATABASE 69
Hatanaka, M., Suzuki, Y., Kawasaki, T. & Endo, M. (1988). Cyclic modified hyperbolic model for sands using pressuremeter test
undrained shear properties of high quality undisturbed Tokyo data. Proc. 5th Int. Symp. on Deformation Characteristics of
gravel. Soils Found. 28, No. 4, 57–68. Geomaterials, Seoul, 949–956.
Ishihara, K. (1996). Soil behaviour in earthquake geotechnics. Park, C.-S. (1993). Deformation and strength characteristics of a
Oxford, UK: Oxford University Press. variety of sands by plane strain compression tests. DEng thesis,
Ito, K., Goto, Y., Ishihara, K., Yasuda, S. & Yoshida, N. (1999). University of Tokyo, Japan (in Japanese).
Detailed in-situ and laboratory tests on the improved ground in Porovic, E. & Jardine, R. J. (1994). Some observations on the static and
Port Island. Proc. 2nd Int. Conf. on Earthquake Geotech. Engng, dynamic shear stiffness of Ham River sand. Proc. 1st Int. Symp. on
Lisboa, 47–52. Pre-failure Deformation of Geomaterials, Sapporo 1, 25–30.
Iwasaki, T. & Tatsuoka, F. (1977). Effects of grain size and grading Roesler, S. K. (1979). Anisotropic shear modulus due to stress
on dynamic shear moduli of sands. Soils Found. 17, No. 3, 19–35. anisotropy. J. Geotech. Engng Div. ASCE 105, No. 7, 871–880.
Iwasaki, T., Tatsuoka, F. & Takagi, Y. (1978). Shear moduli of Rollins, K. M., Evans, M. D., Diehl, N. B. & Daily, W. D. (1998).
sands under cyclic torsional shear loading. Soils Found. 18, No. Shear modulus and damping relationships for gravels. J. Geo-
1, 39–50. tech. Geoenviron. Engng 124, No. 5, 396–405.
Jardine, R. J., Potts, D. M., Fourie, A. B. & Burland, J. B. (1986). Saxena, S. K., Avramidis, A. S. & Reddy, K. R. (1988). Dynamic
Studies of the influence of non-linear stress–strain character- moduli and damping ratios for cemented sands at low strains.
istics in soil-structure interaction. Géotechnique 36, No. 3, 377– Can. Geotech. J. 25, No. 2, 353–368.
396, http://dx.doi.org/10.1680/geot.1986.36.3.377. Seed, H. B. & Idriss, I. M. (1970). Soil moduli and damping factors
Jardine, R. J., Kuwano, R., Zdravkovic, L. & Thornton, C. (2001). for dynamic response analyses, Report EERC 70–10. Berkeley,
Some fundamental aspects of the pre-failure behaviour of gran- CA, USA: University of California.
ular soils. Proc. 2nd Int. Symp. on Pre-failure Deformation Seed, H. B., Wong, R. T., Idriss, I. M. & Tokimatsu, K. (1986).
Characteristics of Geomaterials, Torino, 1077–1111. Moduli and damping factors for dynamic analyses of cohesion-
Jovicic, V. & Coop, M. R. (1997). Stiffness of coarse-grained soils less soil. J. Geotech. Engng 112, No. GT11, 1016–1032.
at small strains. Géotechnique 47, No. 3, 545–561, http:// Shibuya, S., Mitachi, T., Yamashita, S. & Tanaka, H. (1996). Recent
dx.doi.org/10.1680/geot.1997.47.3.545. Japanese practice for investigating elastic stiffness of ground. In
Katayama, I., Fukui, F., Satoh, M., Makihara, Y. & Tokimatsu, K. Advances in site investigation practice (ed. C. Craig), pp. 875–
(1986). Comparison of dynamic soil properties between undis- 886. London, UK: Thomas Telford.
turbed and disturbed dense sand samples. Proc. 21st Ann. Conv. Silver, M. L. & Seed, H. B. (1971). Deformation characteristics of
Japan. Soc. Soil Mech. Found. Engng, Sapporo, 583–584 (in sands under cyclic loading. J. Soil Mech. Found. Div. ASCE 97,
Japanese). No. SM8, 1081–1098.
Koga, Y., Matsuo, O. & Sugawara, N. (1994). In-situ measurement Simpson, B. (1992). Retaining structures: displacement and design.
of shear moduli of soils and its evaluation. In Pre-failure Géotechnique 42, No. 4, 541–576, http://dx.doi.org/10.1680/
deformation of geomaterials (eds S. Shibuya, T. Mitachi and S. geot.1992.42.4.541.
Miura), pp. 213–216. Rotterdam, the Netherlands: Balkema, Skempton, A. W. (1986). Standard penetration test procedures and
Kokusho, T. (1980). Cyclic triaxial test of dynamic soil properties the effects in sands of overburden pressure, relative density,
for wide strain range. Soils Found. 20, No. 2, 45–60. particle size, ageing and overconsolidation. Géotechnique 36,
Kokusho, T. & Esashi, Y. (1981). Cyclic triaxial test on sands and No. 3, 425–447, http://dx.doi.org/10.1680/geot.1986.36.3.425.
coarse materials. Proc. 10th Int. Conf. Soil Mech. Found. Engng, Stokoe, K. H., Hwang, S. K., Lee, J. N. K. & Andrus, R. D.
Stockholm 1, 673–676. (1994). Effects of various parameters on the stiffness and
Kokusho, T. & Tanaka, Y. (1994). Dynamic properties of gravel damping of soils at small to medium strains. Proc. 1st Int.
layers investigated by in-situ freezing sampling. In Ground fail- Symp. on Pre-failure Deformation of Geomaterials, Sapporo 2,
ure under seismic conditions (eds S. Prakash and P. Dakoulas), 785–816.
Geotechnical Special Publication No. 44, pp. 121–140. Reston, Stokoe, K. H., Darendeli, M. B., Andrus, R. D. & Brown, L. T.
VA, USA: ASCE. (1999). Dynamic soil properties: laboratory, field and correction
Laird, J. P. & Stokoe, K. H. (1993). Dynamic properties of remolded studies. Proc. 2nd Int. Conf. on Earthquake Geotech. Engng,
and undisturbed soil samples tested at high confining pressures, Lisbon 3, 811–845.
Geotech. Engrg. Rep. GR93–6. Palo Alto, CA, USA: Electrical Tanaka, Y., Kudo, Y., Yoshida, Y. & Ikemi, M. A. (1987). Study on
Power Research Institute. the mechanical properties of sandy gravel-dynamic properties of
Lanzo, G., Vucetic, M. & Doroudian, M. (1997). Reduction of reconstituted sample, Report U87019. Tokyo, Japan: Central
shear modulus at small strains in simple shear. J. Geotech. Research Institute of Electric Power Industry.
Geoenviron. Engng 123, No. 11, 1035–1042. Tatsuoka, F. & Shibuya, S. (1991). Deformation characteristics of
Lo Presti, D. C. F., Jamiolkowski, M., Pallara, O., Cavallaro, A. & soils and rocks from field and laboratory tests. Proc. 9th Asian
Pedroni, S. (1997). Shear modulus and damping of soils. Géo- Regional Conf. on Soil Mech. and Found. Engng, Bangkok 2,
technique 47, No. 3, 603–617, http://dx.doi.org/10.1680/ 101–170.
geot.1997.47.3.603. Teachavorasinskun, S., Shibuya, S. & Tatsuoka, F. (1991). Stiffness
Mair, R. J. (1993). Developments in geotechnical engineering re- of sands in monotonic and cyclic loading in simple shear.
search: application to tunnels and deep excavations. Unwin Proceedings of the ASCE Geotechnical Engineering Congress,
Memorial Lecture 1992. Proc. Instn Civil Engrs Civ. Engng 97, Boulder, CO. Geotechnical Special Publication No. 27, Vol. 1,
No. 1, 27–41. pp. 863–878. Reston, VA, USA: ASCE.
Maher, M. H., Ro, K. S. & Welsh, J. P. (1994). High strain dynamic Teachavorasinskun, S., Tatsuoka, F., Kenkyo, K. & Yasuhara, K.
modulus and damping of chemically grouted sand. Soil Dynam. (1992). Effect of cyclic prestraining on the liquefaction strength
Earthquake Engng 13, No. 1, 131–138. of sand. Proceedings of the conference on behaviour of offshore
McDowell, G. R. & Bolton, M. D. (1999). A micro mechanical structures, London, pp. 1345–1356.
model for isotropic cyclic loading of isotropically clastically Tika, T., Kallioglou, P. & Pitilakis, K. (1999). Laboratory measure-
compressed soil. Granular Matter 1, No. 4, 183–194. ment of dynamic properties of natural soils. Proc. 2nd Int.
Menq, F. Y. (2003). Dynamic properties of sandy and gravelly soils. Symp. on Pre-failure Deformation Characteristics of Geomater-
PhD dissertation, University of Texas at Austin, TX, USA. ials, Torino 1, 239–244.
Menq, F. Y. & Stokoe, K. H. (2003). Linear dynamic properties of Tika, T., Kallioglou, P., Papadopoulou, A. & Pitilakis, K. (2003).
sandy and gravelly soils from large-scale resonant tests. In Shear modulus and damping of natural sands. In Deformation
Deformation characteristics of geomaterials (eds H. Di Benedet- characteristics of geomaterials (eds H. Di Benedetto, H. Geof-
to, H. Geoffroy, T. Doanh and C. Sauzéat), pp. 63–71. Lisse, froy, T. Doanh and C. Sauzéat), pp. 401–407. Lisse, the Nether-
the Netherlands: Swets & Zeitlinger. lands: Swets & Zeitlinger.
Ogata, N. & Yasuda, M. (1982). Dynamic properties of undisturbed Tokimatsu, K. & Hosaka, Y. (1986). Effects of sample disturbance
samples containing gravels. Proc. 17th Japan. Nat. Conf. Soil on dynamic properties of sand. Soils Found. 26, No. 1, 53–64.
Mech. Found. Engng, Okinawa 1, 1609–1612 (in Japanese). Tokimatsu, K., Yamazaki, T. & Yoshimi, Y. (1986). Soil liquefaction
Oztoprak, S. & Bolton, M. D. (2011). Parameter calibration of a evaluations by elastic shear moduli. Soils Found. 26, No. 1, 25–35.
70 OZTOPRAK AND BOLTON
Vinale, F., d’Onofrio, A., Mancuso, C. & Santucci de Magistris, F. Yang, B. E. (2007). Shear strength modeling of cemented sand,
(1999). The pre-failure behaviour of soils as construction mater- Master thesis, University of Notre Dame, Indiana, USA.
ials. Proc. 2nd Int. Symp. on Pre-failure Deformation Character- Yasuda, N. (1992). Behavior of embankment dams during earth-
istics of Geomaterials, Torino 2, 955–1003. quakes and dynamic deformation characteristics of rockfill mate-
Yamashita, S. & Suzuki, T. (1999). Young’s and shear moduli under rials. Journal Japan Soc. Dam Engs, No. 6, 43–59 (in
different principal stress directions of sand. Proc. 2nd Int. Symp. Japanese).
on Pre-failure Deformation Characteristics of Geomaterials, Yasuda, N. & Matsumoto, N. (1993). Dynamic deformation charac-
Torino 1, 149–158. teristics of sands and rockfill materials. Can. Geotech. J. 30,
Yamashita, S. & Toki, S. (1994). Cyclic deformation character- No. 5, 747–0757.
istics of sands in triaxial and torsional tests. Proc. 1st Int. Yasuda, N. & Matsumoto, N. (1994). Comparisons of deformation
Symp. on Pre-failure Deformation of Geomaterials, Sapporo 1, characteristics of rockfill materials using monotonic and cyclic
31–360. loading laboratory tests and in situ tests. Can. Geotech. J. 31,
Yamashita, S., Shibuya, S. & Tanaka, H. (1997). A case study for No. 2, 162–174.
characterizing undrained cyclic deformation properties in young Yasuda, N., Ohta, N. & Nakamura, A. (1996). Dynamic deformation
sand deposit from in-situ and laboratory tests. Soils Found. 37, characteristics of undisturbed riverbed gravels. Can. Geotech. J.
No. 2, 117–126. 33, No. 2, 237–249.