The Relation Between Liquefaction Resistance and Shear Wave Velocity For New and Old Deposits
The Relation Between Liquefaction Resistance and Shear Wave Velocity For New and Old Deposits
The Relation Between Liquefaction Resistance and Shear Wave Velocity For New and Old Deposits
www.sciencedirect.com
journal homepage: www.elsevier.com/locate/sandf
Technical Paper
The relation between liquefaction resistance and shear wave velocity for new
and old deposits
Roozbeh Safaeian Amolya,n, Kenji Ishiharab, Huriye Bilsela
a
Department of Civil Engineering, Eastern Mediterranean University, North Cyprus, via Mersin 10, Turkey
b
Chuo University, Research and Development Initiative, Tokyo, Japan
Received 28 July 2015; received in revised form 11 October 2015; accepted 27 January 2016
Available online 4 June 2016
Abstract
Multiple series of cyclic triaxial tests were performed on undisturbed and reconstituted samples of sandy soils obtained from areas of known
liquefaction at the time of the 2011 East Japan Earthquake. In this test scheme, the shear wave velocity was firstly measured and then cyclic loads
were applied to determine the cyclic shear strength. The undisturbed samples were classified into two groups, namely, one from old alluvial
(Pleistocene) deep deposits and the other from near-surface shallow depths which had apparently been disturbed by the liquefaction in the 2011
event. The data thus obtained were plotted in terms of the cyclic strength versus the shear wave velocity, and two curved lines were drawn
through average points in the plot for the two groups of soils, that is, one for the undisturbed soils and the other for the liquefaction-disturbed
soils. It was found that for a given cyclic strength, the shear wave velocity does have the propensity to become larger for the undisturbed samples
from the old deposits in comparison to the undisturbed samples from the seemingly liquefaction-disturbed samples. Similar sets of laboratory
tests were also performed on several sand samples reconstituted to a completely disturbed state. The plots of the test results for these reconstituted
samples also showed a tendency in the relation between the cyclic strength and the shear wave velocity which is similar to that for the
liquefaction-disturbed samples recovered from the in-situ deposits. In order to understand the outcome of the above observations, the ratio
between the cyclic strength and the shear modulus from the shear wave velocity was taken as a parameter to distinguish the two different relations
as mentioned above. This ratio, which might be called the “reference strain” of the “yield strain”, is used to provide an interpretation of the
difference in correlation between the cyclic strength and the shear wave velocity.
& 2016 The Japanese Geotechnical Society. Production and hosting by Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.sandf.2016.04.016
0038-0806/& 2016 The Japanese Geotechnical Society. Production and hosting by Elsevier B.V. All rights reserved.
R.S. Amoly et al. / Soils and Foundations 56 (2016) 506–519 507
acceleration of 454 gal which was recorded at a nearby station. A measurements were not performed for the shear wave velocity
plan view of the failed tailings pond is shown in Fig. 4. A typical in the laboratory tests. Thus, only the Vs-values measured in-
soil profile at the centre of the disposal pond is shown in Fig. 5. A situ by the downhole method were included in the data
detailed account of the soil conditions and the features of the compilation file. The remaining half of the undisturbed
failure is described in a paper by Ishihara et al. (2015). samples was transported to the laboratory of another consulting
Undisturbed samples were secured by means of the triple tube company in Chiba, referred to as Co. B, and tested there.
sampler which is said to be suitable for obtaining fairly good- Details of the tests are in this laboratory are as follows.
quality samples from silty sand deposits of low to medium
density. (1) The undisturbed specimens were put into a split tube with
a rubber membrane tightened inside by a vacuum, and then
put in place in the triaxial equipment. One end of the
3. Measurements of shear wave velocity and cyclic membrane was rolled up to the top cap, while the other end
resistance in laboratory was rolled down to the pedestal while a vacuum was
applied inside of the specimen. The split tube was then
3.1. Test procedure and materials removed. Details of the frozen specimen preparation for
cyclic triaxial tests are described in the paper by Ishihara
Undisturbed samples from the tailings disposal pond con- (1985). Prior to applying a back pressure of 200 kN/m2,
tained a fairy large amount of non-plastic fines up to a carbon dioxide (Co2) and then de-aired water were
maximum of 90%. Thus, the sampling tubes were transported circulated through the specimen to achieve a B-value
directly to the laboratory, and the vinyl chloride tubes were cut exceeding 0.95. Once the specimen was isotropically
crosswise and test specimens extruded. The specimens were consolidated to pressures slightly higher (1.1 times) than
trimmed carefully to a cylindrical shape 5 cm in diameter and the in-situ effective overburden corresponding to the depth
10 cm in height. After consolidating the specimens to a at which the specimen had been obtained, shear wave
confining pressure circa 1.1 times the in-situ effective over- velocity measurements were carried out.
burden, tests were performed using the cyclic triaxial test (2) The triaxial test apparatus is equipped with a device to apply
apparatus. The reason for choosing 1.1 times was simply to torsional impulses (step signals) at its top cap and a receiver
bring the specimens to a normally consolidated state as they at the bottom to pick up the impulses transmitted through the
were in-situ. In this laboratory test scheme, on samples from specimen. By dividing the length of the specimen by the time
the tailings dam, shear wave measurements were not of the transmission of the impulses, which were calculated by
carried out. the first arrival method, the shear wave velocity, Vs, was
The tube samples from the Asahi site contained less than obtained. It is to be noted that these impulses produce only an
30% fines, and thus, were hung vertically overnight at the site infinitesimal strain in the specimen. Then, the phase of this
to drain the excess water. The partially saturated sand in the test can be considered as non-destructive.
tubes was frozen in the field and then transported to the
laboratory of a consulting company in Chiba, referred to as Co. (3) After the shear wave measurement is over, the cyclic axial
A. About half of the samples were tested there, but load is applied as many times as necessary to produce a state
R.S. Amoly et al. / Soils and Foundations 56 (2016) 506–519 509
Fig. 3. (a) Soil profile at sampling site in Asahi (HG-S-1). (b) Soil profile at sampling site in Asahi (SN-S-1) (c) Soil profile at sampling site in Asahi (SN-S-2).
of liquefaction and continued even further until the progres- conceived to have identical mechanical characteristics. Note
sive development of axial strain is achieved up to about 10% that this phase of the tests is apparently destructive. Follow-
in single-amplitude under stress-controlled conditions. The ing the cyclic triaxial tests, index property tests were
above tests were repeated 2–4 times by employing different conducted on all the undisturbed specimens to determine
cyclic stress ratios to determine the cyclic strength, RL, for the density, the limiting void ratios, and the specific gravity.
each of the specimens from the same depth which are The grain size distribution curves are shown in Fig. 6.
510 R.S. Amoly et al. / Soils and Foundations 56 (2016) 506–519
Fig. 5. Cross section of failed tailings dam at Ohya mine and locations of boring logs in 1979 and 2011.
4. Results of tests on undisturbed samples with several associated key parameters. The tests on samples
denoted by S-1, S-4, S-6, S-7, S-8, and S-9 from sites HB-S-1,
4.1. Test results on samples from Asahi sites JG-S-1, and NH-S-1 were conducted at the laboratory of Co. A
in Chiba. Shear wave measurements were not performed in
The results of cyclic loading tests on the undisturbed these laboratory tests, and therefore, only the Vs-values
samples from the Asahi sites are shown in Table 1 together obtained in the field by the downhole method are shown in
R.S. Amoly et al. / Soils and Foundations 56 (2016) 506–519 511
the columns, Vs and Vs1. The value of Vs1 is the shear wave laboratory tests. The cyclic loading tests performed to deter-
velocity normalised to the confining stress of s00 correspond- mine the cyclic strength, RL, were performed 2–4 times by
ing to the effective overburden pressure, using the following varying the amplitude of cyclic axial stress. Each time, the
relation: measured Vs-values were different. Thus, the average value
1=4 among the 2–4 samples is indicated in columns Vs and Vs1 with
s0 ' the symbol “L” in Table 1. The ratio, RLPa/G01, was calculated
V s1 ¼ V s ð1Þ
Pa using the same value for RL but with the Vs1 values obtained in
the field and in the laboratory as well. These are shown in the
where the law of normalisation through the minus quarter root
last column of Table 1 with the symbols “F” and “L”,
is the one proposed by Harden and Richart (1963) based on the
respectively.
comprehensive test results. According to their work, the shear
On the basis of the soil profiles shown in Fig. 3, simple
wave velocity was shown also to be a function of the void
liquefaction analyses were conducted using the method
ratio. However, its dependence upon the void ratio is deemed
stipulated in the Japanese design code for bridge design.
to be small particularly for clean sand, as compared to its
The analyses indicated that the recent man-made fills and
dependence upon the confining stress. Thus, only the effect of
alluvial deposits near the surface did indeed experience
the confining stress is considered in this paper to modify the
liquefaction at the time of the 2011 earthquake. Therefore,
measured shear wave velocity. In the same context, shear
these deposits are deemed to have been disturbed by the
modulus G0 was normalised through the minus square root of
strong shaking. The undisturbed samples recovered from
the confining stress, as follows:
such deposits are denoted by “ℓ” in the third column of
1=2 Table 1. The data for the other samples from deeper deposits
s0 '
G01 ¼ G0 ð2Þ are considered not to have liquefied and are labelled as “n” in
Pa
Table 1.
The values for G0 and G01 thus obtained are also shown in
Table 1. The cyclic stress ratio causing 2.5% single amplitude 4.2. Test results on samples from Ohya site
axial strain in 20 cycles, RL, is shown in the second-from- the-
last column in Table 1. Since the value for RL is the ratio The undisturbed samples from the Ohya tailings dam site
between the amplitude of shear stress sd/2 and confining stress were tested in the laboratory of Co. B in Chiba using the
s00 , the value for RL is multiplied by atmospheric pressure triaxial test apparatus. The test procedure was the same as that
Pa ¼ 1 kg/cm2 ¼ 98 kN/m2 so that it is expressed in the same described in the foregoing section. The results of the tests are
unit as the shear modules. shown in Table 2 in the same fashion as in Table 1. In this
The outcome of the majority of the cyclic loading tests on laboratory test series, no measurement was made for the shear
undisturbed samples conducted at the laboratory of Co. B in wave velocity. Thus, the Vs and Vs1 values, shown in Table 2,
Chiba is also shown in Table 1. The tests were performed by are only those monitored in-situ by the downhole method. As a
first monitoring the shear wave velocity, Vs, and then applying result of the simple analysis, the near-surface deposits were
the cyclic load to the same sample. Thus, in columns Vs and identified to have liquefied to depths of about 10 m during the
Vs1, the results are indicated with prescript “L” which implies 2011 earthquake. Thus, the symbol “ℓ” is given for such
512
Table 1
Undisturbed samples from Asahi.
0
Site Sample number Age Depth (m) Void ratio (e) sv (kpa) N-value N1-value FC(%) Vs (m/s) Vs1 (m/s) G01 (Mpa) RL εay ¼ RL pa=G01
HB-S-1 HB-S-1(S-1) Fillℓ 1.0–3.8 0.673 30 7 9 0.9 F160 L168 L52.2 0.304 L** 5.82 10-4
F216 F86.3 F* 4.48 10-4
HB-S-1(S-4) Asn 7.0–10.9 1.409 87 25 27 1 F240 L189 L65.7 0.282 L 4.29 10-4
F248 F116.0 F 2.43 10-4
HB-S-1(S-7) Asn 15.0–16.9 1.094 150 21 26 10.7 F190 L201 L76.4 0.276 L 3.61 10 4
F172 F55.8 F 4.95 10 4
HB-S-1(S-9) Asn 22.0–23.7 1.166 215 12 8 9.6 F190 L208 L81.8 0.276 L 3.37 10 4
F157 F46.5 F 5.94 10 4
JG-S-1 JG-S-1(S-1) Asn 5.0–8.0 1.068 74 25 29 2.2 F180 L150 L42.9 0.176 L 4.10 10 4
F194 F71.1 F 2.48 10 4
JG-S-1(S-4) Asn 8.0–9.9 1.134 91 22 23 6.6 F150 L168 L53.1 0.268 L 5.05 10 4
F154 F44.5 F 6.02 10 4
Table 2
Undisturbed samples from tailings dam site at Ohya mine (Miyagi Prefecture).
Sampling time Age Depth (m) Void ratio (e) s0v (kpa) N-value N1-value FC (%) Vs (m/s) Vs1 (m/s) G01 (Mpa) RL εay ¼ RL Pa =G01
After the 2011 quake Newℓ 5.0–5.8 0.792 85 3 3.3 66 F140 F146 F40.2 0.18 F4.48 10 4
Newℓ 12.0–12.8 1.203 138 1 0.8 96 F170 F157 F46.5 0.173 F3.72 10 4
Newℓ 5.0–5.9 0.809 88 2 2.1 33 F140 F144 F39.1 0.197 F5.04 10 4
Newℓ 8.7–9.5 1.046 120 1 0.9 96 F150 F143 F38.6 0.179 F4.64 10 4
Newℓ 12.0–12.8 0.922 142 2 1.6 97 F170 F156 F45.9 0.208 F4.53 10 4
Newℓ 13.1–14.0 1.046 160 5 3.6 94 F200 F178 F61.4 0.307 F5.00 10 4
Before the 1977 quake Oldn 5.80–6.55 1.46 106 11 10.5 75 F120 F118 F27.0 0.207 F7.67 10 4
Oldn 18.0–18.8 0.73 226 0 0 89 F190 F155 F46.5 0.238 F5.12 10 4
Oldn 5.0–5.8 0.95 91 4 4.2 87 F120 F123 F29.3 0.268 F9.15 10 4
Oldn 10.0–10.8 1.02 128 3 2.5 92 F240 F226 F98.9 0.231 F2.34 10 4
0
N1 ¼ 1:7=ðsv Þþ 0:7 F: Field value, ℓ: Liquefied deposits, n : unliquefied deposits.
strains.” Its physical implication will be discussed later. It may larger than εay ¼ 3.6 10 4 which is demonstrated in Fig. 7
be seen that the value for εay varies in a wide range, but an for the old and unliquefied deposits.
average value representative of all unliquefied old deposits
may be obtained approximately as εay ¼ 3.6 10 4, as indi- 6. Cyclic strength versus shear wave velocity for
cated in Fig. 7. The line of equal εay thus established in the plot reconstituted samples
of RL and Vs1 is regarded as indicating a correlation between
cyclic resistance RL and Vs1 for the old or aged deposits which 6.1. Cyclic strength versus Vs1 for reconstituted samples from
is expressed as Asahi sand
RL ¼ 0:68 10 5 V 2s1 ð4Þ
After the series of tests were finished for the undisturbed
where γ t /g is assumed equal to 18.5/9.8 ¼ 0.189 kN s /m . 2 4
samples from Asahi, the same samples were once mixed, air-
dried, and reconstituted to test specimens having approximately
5.2. Plots of cyclic strength versus Vs1 for fresh fills or the same density. In this case, the tests were carried out only for
liquefied deposits the samples from the HG-S-1, SN-S-1, and SN-S-2 sites. The
Vs-measurements and cyclic loading tests were performed
The test data in Table 1, pertaining to the undisturbed similarly on these reconstituted samples. The results of the tests
samples from fresh artificial fills and liquefied deposits, are shown in Table 3 and the data are plotted in Fig. 9 in terms
are plotted together in Fig. 8 in the same fashion as in Fig. of RL versus Vs1 with the open triangular symbol “◬”. The line
7. By drawing a curved line through the average data of equal εay ¼ 4.6 10 4, established above in Fig. 8, is quoted
points, the relation between R L and V s1 can be obtained as and displayed for comparison. Although there are some devia-
follows: tions, the constant strain εay-line for the reconstituted samples
appears to approximately coincide with that from the in-situ
RL ¼ 0:9 10 5 V s1 2 ð5Þ
samples disturbed by liquefaction.
This relation is displayed in Fig. 8. In the same context as
mentioned above, this curve is a line of the constant value of 6.2. Cyclic strength versus Vs1 for reconstituted samples from
εay ¼ 4.6 10 4. It is to be noted that for the fresh or the once- Nagoya sand
liquefied deposits, the value for εay as above, is found to be
To mainly examine the effects of the fines content, sandy
soil from a site in Nagoya was sorted out using the 74 μ
serve to separate the fines fraction from the silty sand. Then,
the fines was mixed with the sand to produce silty sand
having fines proportions of 10 and 30%. The artificial
materials thus produced were used to prepare specimens
with relative densities of 50% and 70% and tested in the
same fashion using the same apparatus. The results of the
tests are listed in Table 4 and plotted in Fig. 9 with the solid
reverse triangle symbol “ ” in terms of RL versus Vs1. It is
noted that with an increasing relative density, both RL and
Vs1 tend to increase, as seen from the measured values
plotted in the figure. In addition, no clearly discernible
difference in the plots between samples with zero and 30%
fines content appears. Thus, pending further detailed study,
all the samples reconstituted from Nagoya sand are likely, in
the present study, to exhibit seemingly identical character-
istics, at least in terms of the relation between RL and Vs1.
For reference, a line of equal cyclic strain, εay ¼ 4.6 10 4,
is drawn in the plot in Fig. 9. It is noted that the RL Vs1
relation for the Nagoya sand is practically identical to that
for the Asahi sand.
Table 3
Reconstituted specimens from Asahi sand.
0
Specimens FC (%) e s0 ðkPaÞ emax Dr (%) Vs1 (m/s) G01 (Mpa) RL εay ¼ RL pa
G01
emin
HG-S-1 (2.0–3.0 m) 11.8 0.967 100 1.486 83.3 177 59.1 0.19 3.21 10 4
0.863
HG-S-1 (4.0–5.0 m) 12.1 1.02 100 1.572 85.8 135 34.4 0.21 6.10 10 4
0.929
HG-S-1 (18.0–19.0 m) 43.9 1.053 100 1.803 86 115 24.9 0.12 4.82 10 4
0.931
HG-S-1 (19.0–20.0 m) 40.8 1.047 170 1.581 72 125 29.5 0.13 4.41 10 4
0.839
SN-S-1 (1.0–2.0 m) 13.6 1.152 100 1.597 67.6 149 41.9 0.12 2.86 10 4
0.939
SN-S-1 (25.3–26.1 m) 4.5 0.89 270 1.232 73.3 158 47.1 0.175 3.71 10 4
0.768
SN-S-2 (13.0–14.0 m) 5.3 0.837 150 1.298 86.3 168 53.2 0.18 3.38 10 4
0.764
SN-S-2 (20.1–20.8 m) 0.95 0.695 240 0.967 70.1 173 56.4 0.185 3.28 10 4
0.579
SN-S-2 (20.8–1.8 m) 5 1.073 240 1.46 61.3 105 20.8 0.12 5.77 10 4
0.829
Table 4
Reconstituted specimens of Nagoya sand.
indicated by a0 , b0 , c0 , and d0 in the plot of the cyclic stress ratio with the increase in cyclic stress ratio, is considered as a kind
versus the number of cycles. By repeating similar tests with of non-linear stress–strain relationship pertaining to cyclic
different cyclic stress ratios, a series of points can be obtained loading.
similarly for each axial strain amplitude. By connecting the In the theory of elasto-plasticity, it is customary to approxi-
points of equal axial strain amplitude, a set of curved lines can mately represent the non-linear stress–strain relation in terms
be obtained, as indicated in Fig. 10(a). of bi-linear lines. In this case, one line is an asymptote towards
It is well-known that the cyclic stress ratio causing zero point representing the elastic behaviour and another line is
liquefaction with 100% pore water pressure build-up takes an asymptote towards a large strain bounding an upper limit of
place almost concurrently with the development of an axial stress, that is, the strength.
strain of 2.5% in the single amplitude. Thus, it has been a The same idea may be applied in the case of cyclic loading to
common practice to pay attention to the cyclic stress ratio represent the relation between the cyclic stress ratio, Rc, and
causing a single-amplitude axial strain of εa ¼ 2.5% as being a amplitude of cyclic axial strain, εa, as indicated in Fig. 10(b).
state in which liquefaction is triggered. However, since the line Note that the upper limit is now chosen as the cyclic strength, RL,
of equal axial strain, εa ¼ 2.5%, depends upon the number of which is defined as the cyclic stress ratio required to cause 2.5%
cycles, as shown in the plot in Fig. 10(a), it is necessary to single-amplitude axial strain in 20 cycles of load application. The
specify the number to determine a value for the cyclic stress asymptotic line towards zero point is taken as a straight line with
ratio at liquefaction. It has been customary to adopt 15 or 20 a slope, G01, indicating the elastic behaviour at very small strain
cycles of uniform loading as the number of the cyclic loads which is believed with reason to take the same value irrespective
representative of those typically experienced by large earth- of the number of cycles. Thus, the value for G01 can be
quakes with a magnitude of the order of M ¼ 7.5. In the present considered as being the elastic modulus pertaining to the cyclic
study, the number of cycles, Nc ¼ 20, was taken for determin- loading. These two asymptotic lines are indicated schematically
ing the cyclic stress ratio, RL, at liquefaction. If based on this by BC and AB in Fig. 10(b).
presumption, the cyclic stress ratio causing liquefaction is read Suppose the cyclic stress–strain curve of a sandy soil under
off as RL ¼ 0.15, for instance, in the illustrative diagram shown cyclic loading is enclosed by two straight lines, as above, it is
in Fig. 10(a). now possible to determine the axial strain corresponding to
When the equivalent number of cycles is fixed at Nc ¼ 20, it point B in Fig. 10(b) which is an intersection between the two
becomes possible to read off the values of axial strain, εa, as a asymptotic lines. Thus, this strain may be referred to as a kind
function of the applied cyclic stress ratio, Rc, as illustrated by of “reference strain” or “yield strain” in cyclic loading. In order
points a-b-c-d in Fig. 10(a). It is then possible to establish a to define this strain in a non-dimensional format, the cyclic
plot of Rc versus εa as displayed by points a-b-c-d in Fig. stress ratio multiplied by the atmospheric pressure, pa ¼ 1 kg/
10(b). This curve, indicating the growing amplitude of axial strain cm2 ¼ 98 kN/m2, can be used. Therefore, both G01 and
R.S. Amoly et al. / Soils and Foundations 56 (2016) 506–519 517
the cyclic stress at liquefaction, RL pa, are deemed as the 9. Discussions on cyclic yield strain as related to
variables under cyclic loading conditions. It is in this context liquefaction resistance and shear wave velocity
that the ratio between them was referred to as the “cyclic yield
strain”. The yield strain or reference strain in cyclic loading, εay,
A physical implication of the yield strain in cyclic loading, introduced in the present study, appears to have some features
εay, is illustrated in Fig. 10(c). If the value for εay is large, the which are common to the volumetric threshold strain, γtv,
soil is deemed as a material exhibiting a flatter cyclic stress– proposed by Dobry (1985). It was defined as the strain where
strain curve, which might be referred to as “soft” behaviour. In the change in dilatancy-induced volume begins to manifest
contrast, if the yield strain, εay, is small, the soil is considered itself in dry sand or pore water pressure starts to build up in
as “stiff”. Thus, the cyclic yield strain defined by Eq. (3) can be saturated sand in undrained loading. This aspect was more
used as an index property to indicate the levels of softness or thoroughly studied by Hsu and Vucetic (2004) who performed
stiffness of sandy soils under cyclic loading. It is well-known comprehensive laboratory tests on various sandy soils using a
that cyclic strength RL of a soil from a given deposit tends to cyclic simple shear testing device. They showed that the
increase with increasing relative density, but that the initial volumetric cyclic threshold strain tends to increase with the
shear modulus, G01, also increases at the same time. However, plasticity index of fines from about (1–2) 10 4 for clean
the rate of increase in RL pa, and G01 would be in such a way sands up to (4–10) 10 4 for silty sands containing fines with
that the ratio between them, that is, yield strain εay, stays at a a Plasticity Index value of 30. An overview of the existing test
constant value, if the level of softness or stiffness of a given data, reported in 17 papers, was made by Dobry and Abdoun
sand is identical. The idea of constant cyclic yield strain while (2011) who concluded that the value of γtv for normally
changing RL and G01 is illustrated in Fig. 10(d). consolidated sands lies in the range of (0.5–2.0) 10 4 with a
The idea of using a certain level of strain as an index representative value grouping close to around γtv ¼ 1 10 4.
property was first introduced by Dobry et al. (1980) who The relationship between the pore water pressure build-up
proposed the presence of a threshold strain separating condi- and the cyclic shear strain was studied by Ladd. et al. (1989)
tions between occurrence and no occurrence of the dilatancy- who performed a series of strain-controlled cyclic triaxial tests
induced volume change or pore water pressure build-up in on saturated specimens of Monterey No. 0 sand. It was also
cyclic loading. The cyclic yield strain defined above is similar shown that the pore water pressure begins to build up at the
in its vein to that by Dobry, but its interpretation and the strain amplitude of γ ¼ 1.5εa≒1.0 10 4, which is coincident
manner in which it is used would be different. with the threshold volumetric strain proposed by Dobry
(1985). The results of the data compilation made by Dobry
and Abdoun (2011) for the pore water build-up are reproduced
in Fig. 11, together with the data in a similar vein obtained by
8. Factors affecting softness or rigidness of soils under Bhatia (1980) and Finn (1981). As indicated in Tables 1–4, the
cyclic loading range in cyclic yield strain, εay, obtainedin this study, is (2.0–
10) 10 4, and therefore, the cyclic yield strain in simple
Several factors may be envisaged to exert influence on the shear mode, γy ¼ 1.5εa, is (3.0–15) 10 4. This range is also
softness or stiffness of soils as defined above. One of the indicated in Fig. 11. Looking over the dataplots in Fig. 11, one
factors likely to be associated would be the aging of the soil can realise that in the range of the cyclic yield strain, that is,
deposits. It is reasonable to infer that an in-situ soil having a (3.0–15) 10 4, the pore water pressure rises to values 10–
long history of deposition may show stiffer behaviour against 70% of the initially applied confining pressure. Thus, this
the softer response of a soil having a short history of cyclic yield strain, γy, in the simple shear mode, is considered
deposition. Thus, the effect of the aging of soil deposits may
be reflected on the cyclic yield strain, εay.
Another conceivable factor influencing the reference strain,
εay, would be the amount of fines, Fc, contained in the sands in
question. Sandy soils containing larger amounts of fines are
envisaged to exhibit softer behaviour against the rather stiff
characteristics exhibited by soils with lesser amounts of fines.
The reasoning is similar to the above one, and thus, may be
applicable for providing a physical interpretation for the RL
versus G01 relationship for soils with varying contents of fines.
It is known that, for a given soil at a given state with
identical density and stress history, the value of G01 remains
unchanged, whereas the cyclic strength tends to change
significantly with the number of cycles in question. Therefore,
the cyclic yield strain, εay, could also by a function of the
number of load cycles and ought to be specified properly by Fig. 11. Amplitude of simple shear strain versus pore water pressure build-up
reflecting on the duration of shaking during earthquakes. (Whitman, 1985) (Reproduced from figures by Dobry and Abdoun, 2011).
518 R.S. Amoly et al. / Soils and Foundations 56 (2016) 506–519
Fig. 12. Comparison of liquefaction resistance versus shear wave velocity for
newly deposited and old deposits of sandy soils (Reproduced from figure by
Andrus and Stokoe (2000)).
10. Relationship between liquefaction resistance and shear
wave velocity for fresh and old deposits
11. Conclusions
There are several charts hitherto proposed for the relation-
ship between the cyclic strength and shear wave velocity VS1. It has been known that the cyclic resistance of sandy soils to
These are summarised by Andrus and Stokoe (2000) and are liquefaction is influenced by a number of factors, such as the
reproduced in the diagram in Fig. 12. The two lines proposed number of cycles, the fines content, stress histories, the age of
in the present study are superimposed in Fig. 12. It appears that the depositions, etc. In comparison to these, the factors
the majority of the data sets complied hitherto are those from affecting the shear wave velocity or the shear modulus at
field observations and measurements. In view of this, the small strains have not hitherto been targets of detailed
proposed curve pertaining to old deposits with investigations. This trend appears to be the case because of
εay ¼ 3.6 10 4 seems to show a reasonable level of coin- the relatively smaller degree of influence that the shear
cidence particularly with the relations proposed by Tokimatsu modulus exerts on the damage associated with large deforma-
and Uchida (1990) and Robertson et al. (1992). tions of soils during earthquakes. In light of the above
The interpretation for two different relations, with different tendency, it is with good reason that the relationship between
cyclic yield strains, may be made as follows: the cyclic strength and the shear wave velocity is thought to
emerge mainly from the difference in the factors affecting the
1. For a given cyclic strength, RL, the older deposits will yield cyclic strength.
larger values of shear wave velocity as compared to the new With this in mind, an attempt has been made to examine
deposits. whether or not it is appropriate to make use of a single unique
2. For a given shear wave velocity, the new deposits will give relationship for the cyclic strength of sandy soils as correlated
stronger cyclic resistance, RL, than the old deposits. with the shear wave velocity. As one of the factors, the effects
of aging were investigated in this paper by using a cyclic
This appears to contradict intuitive understanding, but if triaxial test apparatus. Tests were performed on undisturbed
interpreted in the context of cyclic yield strain, the new samples recovered from in-situ deposits of known liquefaction
deposits with that Vs-value are already considered to be in a and also on samples from deposits of known non-liquefaction
state which is potentially stronger. at the time of the 2011 East Japan Earthquake. In parallel to
R.S. Amoly et al. / Soils and Foundations 56 (2016) 506–519 519
this, similar series of tests were carried out on reconstituted high-seismicity regions using shear-wave velocity. J. Geotech. Geoenviron.
samples of several sandy soils. The tests consisted of a non- Eng. 141 (3), 04014112 (1–13).
Dobry, R., Powell, D.J., Yokel, F.Y., Ladd, R.S., 1980. Liquefaction potential
destructive type of measurements of shear wave velocity,
of saturated sand – the stiffness method. In: Proceedings of the 7th World
followed by cyclic loading to cause liquefaction. By taking Conference on Earthquake Engineering, Istanbul, Turkey, vol. 3, pp. 25–
the ratio between the cyclic resistance to cause liquefaction in 32.
20 cycles and the square of the shear wave velocity, what Finn, W.D.L., 1981. Liquefaction potential: developments since 1976. In:
might be called the “cyclic yield strain” or the “cyclic reference Proceedings of the International Conference on Recent Advances in
strain” was defined. This is a kind of threshold shear strain at Geotechnical Earthquake Engineering and Soil Dynamics, vol. II,
pp. 655–681.
which the soils begin to deform and yield significantly in the Harden, B.O., Richart, F.E., 1963. Elastic wave velocities in granular soils. J.
cyclic stress–strain curve. Soil Mech. Found. ASCE 89, 33–65.
As a result of the multiple series of tests, undisturbed Hsu, Chu-Chung, Vucetic, M., 2004. Volumetric threshold shear strain for
samples from old aged alluvial deposits were found to cyclic settlement. J. Geotech. Geoenviron. Eng. ASCE 130, 58–70.
exhibit cyclic yield strain which is lower than that from Liu, N., Mitchell, J., 2006. Influence of nonplastic fines on shear wave
velocity-based assessment of liquefaction. J. Geotech. Geoenviron. Eng.
once-liquefied or artificial reclaimed soil deposits. Thus, the 132 (8), 1091–1097.
relationship between these two quantities, as distinguished Ishihara, K., 1985. Stability of natural deposits during earthquakes. In: Pro-
via the relevant cyclic yield strain, is considered to be a ceedings of the 11th International Conference on Soil Mechanics and
useful means for each kind of soil for assessing the Foundation Engineering, San Francisco, vol. 2, pp. 321–376.
Ishihara, K., Ueno, K., Yamada, S., Yasuda, S., Yoneoka, T., 2015. Breach of
liquefaction resistance of in-situ deposits through measure-
tailings dam in the 2011 earthquake in Japan. Int. J. Soil Dyn. Earthq. Eng.
ments of the shear wave velocity. 68, 3–22.
Kayen, R., Moss, R., Thompson, E., Seed, R., Cetin, K., Kiureghian, A.,
Tanaka, Y., Tokimatsu, K., 2013. Shear-wave velocity–based probabilistic
References and deterministic assessment of seismic soil liquefaction potential. J.
Geotech. Geoenviron. Eng. 139 (3), 407–419.
Ahmadi, M., Akbari Paydar, N., 2014. Requirements for soil-specific correla- Ladd., R.S., Dobry, R., Dutko, P., Yokel, F.Y., Chung, R.M., 1989. Pore water
tion between shear wave velocity and liquefaction resistance of sands. Soil pressure buildup in clean sands because of cyclic straining. ASTM
Dyn. Earthq. Eng. 57, 152–163. Geotech. Test. J. 12 (1), 77–86.
Andrus, R.D., Stokoe II, K.H., 2000. Liquefaction resistance of soils from Robertson, P.K., Woeller, D.J., Finn, W.D.L., 1992. Seismic cone penetration
shear-wave velocity. J. Geotech. Geoenviron. Eng. ASCE 126, 1015–1025. test for evaluating liquefaction potential under cyclic loading. Can.
Baxter, C., Bradshaw, A., Green, R., Wang, J., 2008. Correlation between Geotech. J. 29, 686–695.
cyclic resistance and shear-wave velocity for providence silts. J. Geotechn. Stokoe, L.H. II, Roesset, J.M., Biershwale, J.G., Aouad, M., 1988. Liquefac-
Geoenviron. Eng. 134 (1), 37–46. tion Potential of Sands from Shear Wave Velocity. In: Proceedings of the
Bhatia, S., 1980. The Verification of Relationships for Effective Stress Method 9th World Conference on Earthquake Engineering, vol. III, pp. 213–218.
to Evaluate Liquefaction Potential of Saturated Sands. University of British Tokimatsu, K., Uchida, A., 1990. Correlation between liquefaction resistance
Columbia, Canada. and shear wave velocity, soils and foundations. J. Jpn. Geotech. Soc. 30
Dobry, R., 1985. Unpublished Data Files of R. Dobry. Rensselaer Polytechnic (2), 33–42.
Institute, Troy, NY. Wang, J.-H., Moran, K., Baxter, C.D.P., 2006. Correlation between cyclic
Dobry, R., Abdoun, T., 2011. An investigation into why liquefaction charts resistance ratios of intact and reconstituted offshore saturated sands and
work: a necessary step towards integrating the states of art and practice. In: silts with the same shear wave velocity. J. Geotech. Geoenviron. Eng.
Proceedings of the 5th International Conference on Earthquake Geotechni- ASCE 132 (12), 1574–1580.
cal Engineering. San Tiago, Chile, pp. 13–44. Whitman, R.V. (Ed.), 1985. Liquefaction of Soils during Earthquakes,
Dobry, R., Abdoun, T., Stokoe II, K., Moss, R., Hatton, M., El Ganainy, H., Committee on Earthquake Engineering, National Research Council,
2014. Liquefaction potential of recent fills versus natural sands located in National Academy Press, Washington, D.C.