Soil Dynamics and Earthquake Engineering 61-62 (2014) 164–175

Contents lists available at ScienceDirect

Soil Dynamics and Earthquake Engineering

journal homepage: www.elsevier.com/locate/soildyn

Cyclic response of saturated silts

Selman Sağlam n,1, B.Sadık Bakır
Department of Civil Engineering, Middle East Technical University, Ankara 06800, Turkey

art ic l e i nf o a b s t r a c t

Article history: Softening and strength loss of sands with increasing excess pore water pressure under repeated loads is
Received 15 December 2012 well-known. However, extensive damage to the built environment also occurs at the sites underlain by
Received in revised form fine grained soils during seismic shaking. The primary objective of this study is to investigate the factors
16 August 2013
affecting cyclic behavior of saturated low-plastic silt through laboratory testing. For this purpose, an
Accepted 19 February 2014
Available online 15 March 2014
extensive laboratory testing program including conventional monotonic and cyclic triaxial tests was
carried out over reconstituted silt samples. The effects of the inherent soil properties and the effects of
Keywords: loading characteristics on the cyclic response of saturated low-plastic reconstituted silt samples were
Cyclic triaxial test examined separately. Based on the test results, a model was introduced to estimate the effect of initial
Liquefaction
shear stress on the cyclic response. Besides, liquefaction susceptibility of the samples was examined via
Silt
current liquefaction susceptibility criteria.
& 2014 Elsevier Ltd. All rights reserved.

1. Introduction Hence, a detailed laboratory testing program has been undertaken

Response of saturated silty soils under repeated loads has been
a point of interest particularly in the last decade due to the
occurrence of widespread seismically induced damage to the built
environment during recent large earthquakes over areas underlain
by such soils. Owing to the differences between the mechanisms
dominating the cyclic response of fine grained and sandy soils, the
procedures used to estimate the response under seismic loads are
distinct for the two general soil types. The evaluation of the
response of silt, which comprises the borderline between sand
and clay in gradational order, is somewhat more complicated than
those that can be distinguished as sand or clay [1,2]. Although the
information available in literature is still rather limited on the pore
pressure build-up and the associated degradation of stiffness and
strength of silts under cyclic loads, the response is reported to be
dependent on a number of criteria including stress history,
attributes of loading besides the material characteristics such as
the plasticity index. Yet, the findings are often contradictory
regarding whether the influences of such factors are beneficial
or adverse on the response as well as their extents. Accordingly,
the need is clear for further controlled laboratory studies to
improve the present level of knowledge and to clarify the seismic
behavior of silts as emphasized by Sanin and Wijewickreme [3]
and Boulanger and Idriss [1].
n
Corresponding author. Tel.: þ 90 5058336304.
E-mail addresses: ssaglam@adu.edu.tr, se1man@yahoo.com (S. Sağlam).
1
Present address: Department of Civil Engineering, Adnan Menderes
University, Aytepe, Aydın, Turkey.
with the primary aim of improvement of the database for cyclic
response of silts. The major part of the work consists of cyclic and
monotonic triaxial tests conducted over silt specimens. The testing
program is arranged so as to provide a systematic and controlled
investigation of the factors affecting the behavior of silt on a compara-
tive basis. For this purpose, to be able to eliminate the inherent
variability due to natural deposition process, and to provide control
over sample characteristics, reconstituted specimens are used in the
study. Numerous researchers have reported earlier that the reconstitu-
tion procedure may have a substantial effect on the behavior of sands
[4–6]. However, the information concerning the effect of reconstitu-
tion method on the behavior of silty soils in literature is rather limited.
Accordingly, common reconstitution methods were examined. Taking
into consideration the speed and ease of specimen production and
that the saturated specimens are required, utilization of most of the
reconstitution techniques is restricted. Consequently, slurry deposition
technique appeared to be the most convenient for the reconstitution
of saturated silt.
The undrained shear and deformation behavior of the saturated
silt is investigated through a series of monotonic loading, and
stress controlled cyclic triaxial tests conducted over isotropically
and anisotropically consolidated soil samples. Undrained mono-
tonic tests are performed to identify any conceivable relationship
between monotonic and cyclic response of silt. To provide a
profound understanding of the cyclic behavior on various para-
meters, the specimens are tested under a range of preconsolida-
tion pressures, initial static shear stresses and initial confining
stresses. During sample saturation pore pressure coefficient B of at
least 0.95 is provided in all cases in the study.

http://dx.doi.org/10.1016/j.soildyn.2014.02.011
0267-7261 & 2014 Elsevier Ltd. All rights reserved.
 S. Sağlam, B.Sadık Bakır / Soil Dynamics and Earthquake Engineering 61-62 (2014) 164–175 165

100

90

80

PERCENT FINER THAN D

70

60

50

40

30

20

10

0
0.001 0.01 0.1 1
CLAY SILT SAND
PARTICLE SIZE (mm)

Fig. 1. Grain size distribution of the soil utilized in the study.

2. Properties of the reconstituted material and reconstitution Table 1

process Initial states of stress and loading rates of monotonic tests.

The soil used in the tests was supplied in powdered form from Test s'1c (kPa) s'3c (kPa) Initial p'i (kPa) Rate of εa (%/min) OCR ei
Balad, Iraq. The grain size distribution of the light brown colored ST1 35 35 35 0.1 1 0.74
soil is presented in Fig. 1. The material consisted of 68.5% silt, 4.5% ST2 50 50 50 0.07 1 0.72
clay, and 27% fine sand size particles. The specific gravity is ST3 50 50 50 0.07 2 0.71
Gs ¼2.69, and Atterberg limits are determined as LL ¼31 (by means ST4 50 50 50 0.07 4 0.69
ST5 80 80 80 0.07 1 0.73
of Casagrande method), PL¼ 24 and PI¼ 7. The material is classified
ST6 75 75 75 0.05 1 0.74
as low plasticity silt (ML) according to the Unified Soil Classifica- ST7 100 100 100 0.07 1 0.72
tion System (USCS), and plots adjacent to the A-line on the ST8 100 100 100 1 1 0.71
plasticity chart. ST9 100 100 100 1.4 1 0.71
ST10 80 50 60 0.07 1 0.72
During the process of reconstitution, the material was mixed
ST11 100 50 66.67 0.1 1 0.75
with de-aired water of an amount required to bring the water ST12 100 50 66.67 0.1 1 0.76
content to about 2 to 3 times that of the liquid limit. Hence, the ST13 120 50 73.33 0.1 1 0.77
workability is improved during the mixing process, and conse- ST14 90 60 70 0.07 1 0.73
quently homogeneous slurry is provided. The slurry was then ST15 120 60 80 0.07 1 0.74
ST16 140 50 80 0.1 1 0.74
placed into a box having dimensions of 19.5, 19.5 and 21 cm, inside
ST17 150 50 83.33 0.07 1 0.75
of which was covered with a woolen fabric and provided with ST18 160 50 86.67 0.1 1 0.72
narrow drainage holes on the top and bottom covers. The box was ST19 200 50 100 0.07 1 0.73
then submerged in de-aired water and the slurry was consolidated ST20 120 80 93.33 0.1 1 0.71
ST21 150 80 103.33 0.1 1 0.73
under a vertical pressure of 40 kPa, which was imposed by means
ST22 120 100 106.67 0.1 1 0.69
of a pneumatic piston. Following the consolidation, 16 specimens ST23 150 100 116.67 0.1 1 0.68
could be extruded from the box and the applied consolidation ST24 180 100 126.67 0.1 1 0.70
pressure was sufficient to produce specimens that are able to ST25 200 100 133.33 0.1 1 0.70
stand freely under their own weight. The recovered specimens
were trimmed to the dimensions of 36 mm diameter and 71 mm
height before being tested. The rates are determined utilizing the approach to ensure 95% of
pore pressure stabilization in the specimen as proposed by Bishop
and Hengel [7], and Germaine and Ladd [8].
3. Monotonic triaxial compression tests Initial states of stress and applied loading rates for 25 strain
controlled monotonic triaxial tests conducted on the reconstituted
A series of undrained monotonic triaxial tests were conducted specimens are listed in Table 1. In addition to the tests performed
over the reconstituted samples to examine the response under with loading rates between 0.05 and 0.1%/min, two additional
monotonic loading, and to identify any conceivable relationship tests were carried out at a loading rate of 1.4 and 1%/min on the
between monotonic and cyclic responses. An important issue specimens isotropically consolidated under 100 kPa. Based on the
during triaxial testing of fine grained soils is the rate at which test results, the average values of internal friction angle and
the load is applied. Due to the friction at the two ends, distribution cohesion were determined as 371 and 5 kPa, respectively.
of stress and the strain is non-uniform along the specimen during Two specimens with OCR of 2 and 4 were monotonically tested
loading. Sufficient time should be allowed during loading to as well. State of overconsolidation was achieved through unload-
provide stabilization of the pore water pressure within the speci- ing the specimens that were isotropically consolidated under 100
men. Otherwise, the strength of the soil is affected by the non- and 200 kPa, respectively. Accordingly, both OC specimens sus-
uniformity of pore pressure. The monotonic triaxial tests were tained 50 kPa initial effective confining stress prior to shearing.
conducted at rates of strain ranging between 0.05 and 0.1%/min. Those specimens were tested with a loading rate of 0.07%/min.
166 S. Sağlam, B.Sadık Bakır / Soil Dynamics and Earthquake Engineering 61-62 (2014) 164–175

The initial stress conditions of the specimens were arranged to

provide a wide range of representative in-situ stress states
representative of those that may exist beneath the foundations
as well as at the free field.
Results of the eight monotonic tests are presented in Fig. 2 in
the form of plots of mean effective stress (p') versus deviator stress
(Δs), where p' ¼[s'1 þ 2s'3]/3 and Δs ¼ [s1  s3]. Fig. 2a and b
respectively show the responses of the specimens consolidated
isotropically under different all-round pressures (s'3c), and sus-
taining various initial shear stress ratios (τs/p'i); where, τs is the
static shear stress and p'i is the initial mean effective stress. Fig. 3
displays the excess pore water pressure generation during mono-
tonic shearing. The excess pore pressure generated during shear-
ing is observed to decrease with increasing τs/p'i. As it can be
concluded from Figs. 2 and 3, subsequent to an initial contractive
response, the specimens exhibit dilation upon continued shearing.
The phase transformation starts earlier with increasing τs/p'i.
Although the Figs. 2 and 3 display results of only 8 monotonic
tests, it is to be noted that almost all the specimens eventually
dilate regardless of initial stress state.
Stress–strain responses of the specimens subjected to
undrained monotonic loading with no initial shear stress and with
various initial shear stress ratios are presented in Fig. 4a and b,
correspondingly. As it can be observed, a point of flexure exists at
very low strains (0.1 0.6%) in the stress–strain plots, which is
followed by hardening up to the axial strains of 10% without a
clear peak is being reached. As a general trend, the undrained
strength increases with increasing initial shear stress.
The results of ST2, ST3 and ST4 tests, conducted to investigate Fig. 3. Monotonic pore pressure generation for the specimens with (a) no initial
the influence of overconsolidation on the response to monotonic shear stress, and (b) a range of initial shear stresses.
shearing, are presented in Fig. 5. The specimens in these tests were
imposed a s'3c of 50 kPa prior to shearing. Therefore, there exists
no influence of confining stress on the plotted relationships.
The undrained shear strength is observed to increase with

Fig. 4. Monotonic stress–strain behavior of the specimens with (a) no initial shear
stress, and (b) a range of initial shear stresses.

increasing OCR in Fig. 5a, whereas the pore water pressure

Fig. 2. Monotonic stress paths for the specimens with (a) no initial shear stress, generation decreases with increasing OCR and the pore pressure
and (b) a range of initial shear stresses. takes negative values. Thus, it can be stated that the tendency to
 S. Sağlam, B.Sadık Bakır / Soil Dynamics and Earthquake Engineering 61-62 (2014) 164–175 167

Table 2
Initial conditions of cyclic triaxial tests.

Test s'1c (kPa) s'3c (kPa) τs/p'i CSRtx ei w/LL OCR f (Hz)
(Δscyc/2p'i)

C1 50 50 0 0.35 0.74 0.84 1 0.5

C2 50 50 0 0.3 0.74 0.843 1 0.5
C3 50 50 0 0.55 0.75 0.849 1 0.5
C4 50 50 0 0.6 0.75 0.851 1 0.5
C5 45 45 0 0.72 0.72 0.82 1 0.5
C6 75 50 0.21 0.34 0.74 0.835 1 0.5
C7 75 50 0.21 0.51 0.72 0.812 1 0.5
C8 75 50 0.21 0.69 0.75 0.846 1 0.5
C9 75 50 0.21 0.69 0.75 0.852 1 0.5
C10a 95 50 0.35 0.35 0.75 0.851 1 0.5
C11 95 50 0.35 0.54 0.75 0.848 1 0.5
C12 90 50 0.32 0.71 0.74 0.844 1 0.5
C13a 120 60 0.38 0.41 0.75 0.857 1 0.5
C14a 120 60 0.38 0.5 0.77 0.875 1 0.5
C15 120 60 0.38 0.59 0.74 0.839 1 0.5
C16 80 80 0 0.31 0.75 0.854 1 0.5
C17 80 80 0 0.41 0.74 0.838 1 0.5
C18 80 80 0 0.5 0.75 0.848 1 0.5
C19 80 80 0 0.59 0.74 0.842 1 0.5
C20 80 80 0 0.59 0.73 0.825 1 0.5
C21 80 80 0 0.59 0.73 0.833 1 0.5
C22 90 60 0.21 0.36 0.74 0.841 1 0.5
C23 90 60 0.21 0.54 0.75 0.85 1 0.5
C24 90 60 0.21 0.64 0.72 0.821 1 0.5
C25a 150 50 0.6 0.45 0.75 0.851 1 0.5
C26a 150 50 0.6 0.6 0.74 0.845 1 0.5
C27 50 50 0 0.3 0.73 0.828 2 0.5
C28 50 50 0 0.5 0.73 0.827 2 0.5
Fig. 5. Monotonic (a) stress–strain behavior and (b) pore pressure generation for C29 50 50 0 0.6 0.73 0.825 2 0.5
the specimens with different OCR values. C30 50 50 0 0.6 0.73 0.83 2 0.5
C31 50 50 0 0.65 0.73 0.826 2 0.5
C32 50 50 0 0.35 0.7 0.789 4 0.5
C33 50 50 0 0.35 0.7 0.79 4 0.5
C34 50 50 0 0.5 0.69 0.781 4 0.5
dilate becomes more prevalent with increasing OCR during the
C35 50 50 0 0.55 0.7 0.797 4 0.5
monotonic response. C36 50 50 0 0.6 0.68 0.773 4 0.5
C37 50 50 0 0.8 0.69 0.778 4 0.5
C38 95 95 0 0.45 0.72 0.814 2 0.5
C39 75 75 0 0.47 0.72 0.822 2.5 0.5
4. Cyclic triaxial tests
C40 50 50 0 0.55 0.72 0.817 2 0.5
C41 100 100 0 0.55 0.7 0.797 2 0.5
Total of 69 cyclic tests were conducted over isotropically and C42 100 100 0 0.48 0.72 0.813 1.5 0.5
anisotropically consolidated silt specimens. The consolidation pres- C43 100 100 0 0.3 0.7 0.796 2 0.5
C44a 95 50 0.35 0.4 0.76 0.858 1 0.5
sures were intended to be representative of those likely to exist over
C45 90 60 0.21 0.48 0.73 0.829 1 0.5
the soil elements within the depths of interest beneath foundations. C46 90 60 0.21 0.61 0.74 0.837 1 0.5
The range of cyclic stress ratios (CSR) were selected so as to commonly C47a 150 50 0.6 0.7 0.74 0.838 1 0.5
represent the amplitudes of loading during seismic events. Accord- C48a 200 50 0.75 0.39 0.73 0.829 1 0.5
ingly, a range of CSRtx values between 0.30 and 0.72 were utilized, and C49a 195 50 0.74 0.6 0.74 0.841 1 0.5
C50 100 100 0 0.34 0.74 0.843 1 0.5
the corresponding amplitudes of cyclic stress were applied in a stress- C51 75 50 0.21 0.4 0.72 0.817 1 0.5
controlled manner. The ranges of stress and loading frequency; stress C52a 95 50 0.35 0.42 0.76 0.858 1 0.5
states previous to the cyclic phase; void ratio and the ratio of water C53 50 50 0 0.5 0.74 0.843 3 0.5
content to LL at the beginning of cyclic shearing for the tested samples C54 50 50 0 0.62 0.72 0.82 4 0.5
C55 50 50 0 0.7 0.72 0.82 3 0.5
are presented in Table 2.
C56 50 50 0 0.53 0.72 0.812 4 0.5
Majority of the tests were conducted at a frequency of 0.5 Hz, C57 50 50 0 0.52 0.74 0.843 2 0.5
which is considered as representative of the typical frequency range C58 50 50 0 0.58 0.74 0.843 2 0.5
of seismically induced loading. In order to examine the influence of C59 50 50 0 0.53 0.72 0.812 1 0.05
C60 50 50 0 0.85 0.71 0.804 1 0.05
loading rate on the response, 11 of the tests were carried out at a
C61 120 60 0.38 0.54 0.72 0.812 1 0.05
frequency of 0.05 Hz. In order to investigate the influences of stress C62 80 80 0 0.54 0.74 0.843 1 0.05
history and initial stress state on the cyclic response of silt, the tests C63 150 50 0.6 0.6 0.73 0.827 1 0.05
were carried out with various values of OCR and initial shear stress C64 50 50 0 0.3 0.72 0.82 1 0.05
ratios (τs/p'i). To impose the target OCR, following the initial isotropic C65 100 50 0.38 0.37 0.72 0.82 1 0.05
C66 120 60 0.38 0.44 0.72 0.82 1 0.05
consolidation phase under a specific confining stress, the drainage C67 80 80 0 0.44 0.73 0.827 1 0.05
valve is closed and the confining pressure was appropriately C68 150 50 0.6 0.48 0.71 0.804 1 0.05
reduced. Thus, the OC specimens had isotropic initial stress states C69 100 50 0.38 0.56 0.71 0.804 1 0.05
ahead of being subjected to cyclic loading. On the other hand, the a
Number of cycles was calculated considering SA axial strains.
specimens in NC state were consolidated under various ratios of
initial shear stress to initial mean effective stress (τs/p'i). The applied consolidation. The sustained initial shear stresses were provided
initial shear stress ratios over the specimens range between nil and through increasing the deviatoric stresses (Δs) after consolidation
0.75. It should be noted that τs/p'i ¼0 condition refers to the isotropic under confining pressures (s'3c). Then, consolidation was continued
168 S. Sağlam, B.Sadık Bakır / Soil Dynamics and Earthquake Engineering 61-62 (2014) 164–175

Fig. 6. Stress–strain behavior for the case of non-reversal stress conditions

observed for the tests of (a) C10, where CSRtx ¼ 0.35 and τs/p'i ¼ 0.31; and (b) C13,
where CSRtx ¼0.41 and τs/p'i ¼ 0.33.

under increased axial stress (s'1c ¼ s'3c þ Δs) and s'3 for a sufficient
period of time.
Cyclic response is observed to depend significantly on whether
the specimens are subjected to stress reversal or not during cyclic
loading. In the case of no stress reversal, plastic strains accumulate
with almost a constant rate in each cycle. As it can be seen in Fig. 6,
the plastic strain accumulation rate tends to decrease after the
peak cyclic stress remains below the monotonic strength. The
greater the ratio of the applied peak cyclic stress to the monotonic
strength, the strains occur with a greater rate of accumulation,
which is consistent with the observed trends reported in literature
[9,10]. No significant cyclic degradation was observed in the
stiffness of the specimens when loaded without stress reversals,
although the axial strains exceeded 5%.
Fig. 7 shows the results of the tests with the isotropically consoli-
dated specimens subjected to two-way loading under different CSRtx
and p'i values. The axial strain accumulation rate is observed to Fig. 7. Stress–strain behavior of the specimens subjected to (a) CSRtx of 0.31 and p'i
increase for CSRtx value of 0.31. There is no sudden increase in the of 80 kPa (C16), (b) CSRtx of 0.50 and p'i of 80 kPa (C18).
strain to be interpreted as flow liquefaction. At relatively high CSRtx of
0.5, strain accumulation rate decreases with increasing number of
cycles N although the strain accumulates particularly in the earlier monotonic strength as well as with the increasing CSRtx. The pore
cycles of loading. In most of the isotropically consolidated cyclic tests, a water pressure ratio ru generated in the specimen having an initial
loading rate of 0.5 Hz was used. The ru values, were, however, shear stress ratio τs/p'i of 0.21, approaches to 0.8, whereas it
observed to reach, and even exceed 0.9 at different cycles. Although approaches to 0.5 for τs/p'i of about 0.35. However, majority of DA
the excess pore pressure reaching at the initial confining stress led to axial strains of the tests exceed 10% regardless of ru value.
the loss in effective stress and to cyclic strain softening as a
consequence, the tendency of dilation of the silt prevented excessive
loss of strength. 4.1. Effect of initial confining stress
In those tests in which the anisotropically consolidated speci-
mens subjected to increased loading demands (higher CSRtx In free field, soil elements are by and large subjected to uneven
values), the stress reversals also occurred. Fig. 8 shows the test horizontal and vertical stresses. However, as indicated by Seed and
results of anisotropically consolidated specimens subjected to Lee [11], the stress condition initially existing in free field is
stress reversals. The strains predominantly accumulate on the simulated through application of isotropic confining stress. There-
compression side of loading, and the accumulation rate increases fore, it is also of concern to distinguish the effects of initial shear
with the increasing extent between the peak stress and the stress and the effects induced by the confining stress.
 S. Sağlam, B.Sadık Bakır / Soil Dynamics and Earthquake Engineering 61-62 (2014) 164–175 169

Fig. 9. Relationship between DA axial strain and N for reconstituted NC specimens

conducted with CSRtx of (a) 0.30 and (b) 0.60.

Fig. 8. Stress–strain behavior of the specimens subjected to (a) CSRtx of 0.51

and τs/p'i of 0.21 (C7), (b) CSRtx of 0.54 and τs/p'i of 0.35 (C11).

Isotropic consolidation phase during triaxial tests were carried

out mostly with confining pressures of 50, 80 and 100 kPa. Fig. 10. Relationship between CSRtx and N to reach at 5% DA axial strain for
The effect of confining stress on the relationship between DA axial reconstituted NC specimens consolidated under s'3c of 50 and 80 kPa.
strain and number of cycles (N) can be observed in Fig. 9. The
responses presented in Fig. 9a belong to the tests performed with
NC specimens. The specimens in the tests C2, C16 and C50 were
consolidated under isotropic confining stresses of 50, 80 and
100 kPa, respectively, and were loaded with a CSRtx of 0.30. As it
can be inferred from the figure, cyclic straining tends to increase in
the earlier cycles with increasing confining stress. In Fig. 9b, the
responses obtained in tests C4 and C19, which were consolidated
under s'3c of 50 and 80 kPa, are given. The CSRtx used in these
tests was 0.60. Even if the decreasing resistance is not as much
clear as that for the case of CSRtx of 0.30, the cyclic resistance also
decreases with increasing confining stress.
The influence of confining stress on the cyclic response of NC
specimens is also shown by the plots given in Fig. 10. The plots
Fig. 11. Relationship between DA axial strain and N for reconstituted specimens
show the relationship between CSRtx and N to reach 5% DA axial having OCR of 2 that were loaded with CSRtx of 0.55.
strain for the tests conducted with s'3c of 50 and 80 kPa. It is seen
that N required to reach 5% DA axial strain is greater for the tests
conducted with s'3c of 50 kPa compared with those conducted two OC specimens tested with CSRtx of 0.55 are illustrated in
with s'3c of 80 kPa. Fig. 11 as a function of N. The tests C40 and C41 with OCR of 2 were
The effect of initial confining stress on cyclic response of NC consolidated under s'3c of 50 and 100 kPa, respectively. The
reconstituted silt is observed to be parallel with those reported in specimen tested with s'3c of 100 kPa undergoes high strains
the literature [12–14]. Nevertheless, this effect is examined for the earlier than that tested with s'3c of 50 kPa. In Fig. 12, CSRtx is seen
lightly OC silt as well in the present study. The strain responses of to be greater for the specimens consolidated under s'3c of 50 kPa.
170 S. Sağlam, B.Sadık Bakır / Soil Dynamics and Earthquake Engineering 61-62 (2014) 164–175

The tests conducted with frequency of 0.5 Hz are considered

Fig. 12. Relationship between CSRtx and N to reach at 5% DA axial strain for
reconstituted specimens with OCR of 2 that were consolidated under s'3c of 50 and
100 kPa.
Thus, it can be concluded that the increasing confining stress has a
crucial effect on the cyclic resistance of reconstituted NC and OC
silt with OCR of 2. The tests performed with specimens having OCR
values greater than 2 were carried out with only s'3c of 50 kPa,
therefore the effect of confining pressure of specimens with OCR
greater than 2 cannot be evaluated. Nevertheless, it must be
pointed out that Voznesensky and Nordal [15] showed that the
cyclic resistance of clays increases with increasing confining stress
for the samples in OC state, whereas the opposite is observed for
NC samples.
4.2. Effect of initial shear stress
here for examination of the effect of initial shear stress. In the
cyclic tests, the accumulation of axial strains mostly reaches or
exceeds 10%. In some of the tests, however, the cyclic strain
accumulation levels remains below a certain value, and even
below 5%. Therefore, as an optimal value reached overall in the
tests εa of 5% is chosen as the reference strain when examining the
initial shear stress effect. As it can be observed in Fig. 13, N
required reaching εa of 5% increases with increasing τs/p'i for a
given value of CSRtx. Nevertheless, the specimens having τs/p'i of
0.75 display a reduced cyclic strength compared to those having
τs/p'i of 0.60. Although the curve for τs/p'i of 0.75 is plotted based
on only two tests, it is seen that the initial shear stress exceeding
60–70% of p'i causes a decrease in the cyclic strength of the silt.
The observed tendency of cyclic strength to increase with increas-
ing initial shear stress is identical with the data for sands having
relative densities greater than 50%, as reported by Ishihara [13].
The increase in strength becomes more evident when the results
are interpreted in terms of maximum total undrained shear stress
ratio ((τs þ τcyc)/p'i). Fig. 14 shows that the total shear stress ratio
required to reach an axial strain of 5% increases with a given N. The
specimens having τs/p'i of 0.75 are observed to have a greater total
undrained shear stress ratio than those having 0.6.
The trend observed in the tests conducted in this study is
contradictory to that reported by Hyde et al. [20]. The cyclic
resistance is found to increase more apparently for the tests with
stress reversal, and no significant influence of increasing initial
shear stress is detected for the case of no stress reversal.
The curves fitted to the observed data in Fig. 13, correlating CSRtx
to values of N required to reach at 5% εa, are given by

The soil elements beneath existing foundations or in slopes are CSRtx ¼ c  N d ð1Þ
subjected to initial shear stresses previous to a potential cyclic
loading. In order to simulate the in-situ stress states of such
elements, the specimens are anisotropically consolidated pre-
viously in cyclic triaxial testing. Cyclic strength of fine grained
soils has been reported to display both decreasing and increasing
trends with increasing initial shear stress. Andersen et al. [9] and
Yasuhara et al. [16] reported that the imposed initial shear stress
increases the cyclic shear strength of normally consolidated clays.
For the overconsolidated clays, Andersen et al. [9] reported that
the cyclic strength appears to be independent of the initial shear
stress. Konrad and Wagg [17] also observed the increasing trend of
cyclic resistance with increasing initial shear stress for 20% clay
and 80% silt mixture. Hyodo et al. [18] stated that the cyclic
strength of NC clays decrease with increasing initial shear stress.
Interesting point of that study is that the cyclic strength of sands
with relative densities of 50 and 70% are compared to that of clay Fig. 13. The relationship between CSRtx and N to reach at 5% εa for different τs/p'i
with reference to the sustained initial shear stresses. The cyclic values.
strength of clay is almost twice of that of sands for isotropically
consolidated states of stress, whereas the cyclic strength of clay is
reduced well below those of sands with the increasing initial shear
stress. Ishihara [13] showed that if the applied initial shear stress
exceeds 90% of the static strength of clay, the cyclic strength is
prone to severe decrease with increasing initial shear stress. On
the other hand, an initial shear stress ranging between 50 and 80%
of the static strength is reported as not having a significant
influence on the cyclic strength of the clay tested in the study.
The same tendency is also observed by Lefebvre and Pfendler [19]
for soft clay specimens with initial shear stresses ranging between
nil and 80% of the undrained static strength. Hyde et al. [20]
reported that the increasing initial shear stress decreases the cyclic
resistance of the silt specimens subjected to stress reversals during
cycling, whereas the cyclic resistance of the silt specimens displays
an increasing trend with the increasing initial deviator stress for Fig. 14. The relationship between (τs þτcyc)/p'i and N to reach at 5% εa for different
no stress reversal case. τs/p'i values.
 S. Sağlam, B.Sadık Bakır / Soil Dynamics and Earthquake Engineering 61-62 (2014) 164–175
Fig. 15. The model representing the effect of initial shear stress on cyclic response
of reconstituted silt.
where the coefficients c and d are expressed in terms of
values as follows:
c ¼ 0:7τs =p'i þ 0:5
d ¼  0:18ðc þ 0:1Þ
The relationship modeled by the correlation presented above is
plotted in Fig. 15 in the form of curves corresponding to different
values of N. The model appears to be in good agreement with the
observed effect of initial shear stress on cyclic response of
reconstituted silt.
4.3. Effect of overconsolidation ratio
Stress history of the soils is an important factor influencing
stress–strain response under both monotonic and dynamic load-
ing. This effect is more obvious for saturated fine grained soils,
since the fine grained soil behavior is intrinsically dependent on
the stress history. Although increasing OCR has generally been
reported as causing an increase in the monotonic strength of fine
grained soils, there exist studies with contradictory finding
regarding the cyclic strength [9].
In this study, OC states were attained by modification of s'3c
following an isotropic consolidation phase. In the tests OC states
were obtained typically by reloading to s'3c of 50 kPa. Thus, the
specimens with OCR of 1, 2, 3 and 4 were initially consolidated
under 50, 100, 150 and 200 kPa, respectively. Accordingly, effect of
OCR on the monotonic and cyclic behavior of reconstituted silt
would be investigated for the specimens sustaining no initial shear
stress.
The effect of OCR on the cyclic response is shown in Fig. 16 via
relationship between CSRtx and N required to reach at 5% εa. As it
can be observed, CSRtx that would cause the value of N required
reaching at 5% εa increases with increasing OCR. However, the
increase in resistance is not as significant as that observed in the
monotonic tests. Even for some specific tests, it is observed that
increasing OCR causes a decrease in N required to reach 5% εa.
The excess pore water pressure generation observed during
cyclic tests is displayed in Fig. 17. As it is seen, the maximum value
of ru in each cycle decreases with increasing OCR. Interestingly,
however, certain axial strains are reached at lower excess pore
water pressures with increasing OCR. Nonetheless, if the cyclic
resistance is evaluated by means of axial strain accumulation, OCR
effect on cyclic resistance is not that clear as observed in mono-
tonic tests. On the other hand, if the cyclic resistance is evaluated
by means of pore water pressure generation, increase in the cyclic
resistance becomes more obvious with increasing OCR.
171
Fig. 16. The relationship between CSRtx and N to reach at 5% εa for different OCR
values.
τs/p'i
ð2Þ
ð3Þ
Fig. 17. Excess pore pressure generation with axial strains for different OCR levels.
4.4. Effect of loading rate
Loading rate is another factor that has been frequently pro-
nounced in literature to have a significant effect on the stress–strain
behavior, particularly of the fine soils. The strain developing during
cycling significantly depends on the loading amplitude (i.e., CSR),
and the strain developed in a certain time interval is the indicator of
the loading rate. Hence, development of strain is based on both, the
loading frequency and the loading magnitude for cyclic tests. The
effect of loading rate on cyclic behavior is examined via tests
conducted under frequencies of 0.5 and 0.05 Hz. In order to discard
the loading magnitude component, conjugate tests (i.e. loaded
under similar CSRtx values) conducted with different frequencies
are to be compared.
The influence induced by loading rate has been typically
attributed to the pore pressure generation. Due to the rather low
hydraulic conductivities of fine grained soils, the pore water
pressure within the specimen would not be uniformly distributed
under high speed loading. The excess pore water pressure gen-
eration during cyclic loading can be taken into consideration as
separate components of transient and residual excess pore pres-
sures [21]. The residual component is described as the excess pore
water pressure built up at the state of zero cyclic deviator stress in
a cycle while the transient one is the offset of the excess pore
pressure around residual excess pore water pressure in a cycle.
During cycling, the residual pore water pressure increases with
increasing number of cycles, and the transient pore water pressure
becomes relatively negligible compared to the residual. Hence, the
equalization of residual pore water pressure continues throughout
the cycling process and improves with increasing number of
cycles. Accordingly, the loading rate can be stated to influence
the pore pressure response rather in the early cycles.
The ru and N values at 5% of εa for the conjugate tests
conducted with two different frequencies are shown in Table 3.
The loading rate effect differs depending on the existence of initial
shear stress. It is observed that εa of 5% is reached by the first cycle
for the tests with 0.05 Hz, whereas it is reached at N between 2
172 S. Sağlam, B.Sadık Bakır / Soil Dynamics and Earthquake Engineering 61-62 (2014) 164–175

Table 3
Values of ru and N observed at εa of 5% for conjugate tests conducted with 0.5 and
0.05 Hz.

Conjugate tests s'1c (kPa) s'3c (kPa) τs/p'i CSRtx f (Hz) At 5% εa

ru (%) N

C3 50 50 0 0.55 0.5 68 7
C59 50 50 0 0.53 0.05 51 1
C5 45 45 0 0.72 0.5 73 25
C60 50 50 0 0.85 0.05 35 1
C18 80 80 0 0.5 0.5 25 2
C62 80 80 0 0.54 0.05 55 1
C17 80 80 0 0.41 0.5 51 6
C67 80 80 0 0.44 0.05 40 1
C10a 95 50 0.31 0.35 0.5  
C65a 100 50 0.33 0.37 0.05  
C13a 120 60 0.33 0.41 0.5 71 46
C66a 120 60 0.33 0.44 0.05 76 91
C11 95 50 0.31 0.54 0.5 48 3
C69 100 50 0.33 0.56 0.05 46 3
C15 120 60 0.33 0.59 0.5 30 2
C61 120 60 0.33 0.54 0.05 58 2
C26a 150 50 0.5 0.6 0.5 41 9
C63a 150 50 0.5 0.6 0.05 27 2
C25a 150 50 0.5 0.45 0.5 55 113
C68a 150 50 0.5 0.48 0.05 52 39

a
Number of cycles was calculated considering SA axial strains.

and 25 for the tests with 0.5 Hz when compared to the conjugate
tests carried out with no initial shear stress. Thus, increase in loading
rate obviously causes increase in the cyclic resistance of the silt
specimens that do not sustain initial shear stress. On the other hand,
N values observed at 5% εa for the conjugate tests conducted with
specimens subjected to initial shear stress do not display differences
as significant as those observed for the specimens with no initial
shear stress. N observed for a test with 0.5 Hz is even lower than that
observed in the conjugate test with 0.05 Hz. Observed ru at εa of 5% is
different as well as N in the conjugate tests performed with no initial
shear stress while ru values of the conjugate tests are close to each
other for the specimens in most of the tests carried out with an initial
shear stress. Therefore, in the general sense it can be stated that the
loading rate does not influence the cyclic resistance if the loading
remains on the compression side.

4.5. Cyclic stiffness degradation

Cyclic degradation in soils is measured by means of the

degradation index (δD). The index defines the relative variation
of secant modulus (E) or shear modulus (G) by the ratio of the
modulus in the first cycle (E1 or G1) to the modulus at the Nth cycle
(EN or GN). Idriss et al. [22] and Ishihara [13] showed that δD
decreases linearly with increasing N. The inclination of this
straight line increases with increasing cyclic loading amplitude.
The slope of the line is termed as degradation parameter (t) and
associated to stress history in the case of fine grained soils.
The cyclic degradation observed in the reconstituted silt speci-
mens in this study is quantified based on tests conducted with
isotropically consolidated specimens. The stiffness degradation
diminishes with increasing initial shear stress ratio and becomes
negligible with increasing N for the tests with no stress reversal.
The cyclic degradation is measured here through the ratio of
secant modulus (E) calculated at the first and Nth cycles. The
secant modulus is calculated considering εa observed at peak
Δscyc for the relevant cycle. Then, δD is calculated for each cycle by
considering SA εa values observed at compression side.
To illustrate the cyclic degradation for the silt specimens, the
relationship between δD and N is plotted as a function of the
applied CSRtx in Fig. 18, in which different OC states were
Fig. 18. Stiffness degradation with increasing cycles for the silt specimens with OCR
of (a) 1, (b) 2 and (c) 4.
evaluated with separate plots. As would be expected, the cyclic
degradation is observed to increase with increasing CSRtx for all
OCR values. Increasing CSRtx causes faster degradation of stiffness
in all the specimens. The trendlines fitted to represent the
relationship between δD and N deviate from being linear with
increasing CSRtx, although it was observed and modeled earlier in
the form of a straight line by Idriss et al. [22]. The trendlines
observed for the tests conducted with a CSRtx around 0.30 are
linear for OCR values of 1 and 4. Interestingly, each trendline plots
a curve rather than a line for the specimens having OCR value of 2.
In order to evaluate the effect of stress history on degradation
tendency, the degradation parameter (t) was calculated for the
linear part of the trendlines shown in Fig. 18. The relationships
between t and CSRtx for OCR values of 1, 2 and 4 are given in
Fig. 19. It is observed that an increase in OCR causes reduction in
the stiffness degradation as noted by Vucetic and Dobry [23].This
 S. Sağlam, B.Sadık Bakır / Soil Dynamics and Earthquake Engineering 61-62 (2014) 164–175 173

Andrews and Martin [27] reviewed a number of earthquake

case histories, and discussed the properties of the soils that were
documented as liquefied. Clay content and LL were regarded as
“key” parameters on liquefaction susceptibility evaluation of silty
soils, and it was concluded that the soils are susceptible to
liquefaction if they have LL o32 and clay content o10%, and are
not susceptible if they do not satisfy these two criteria simulta-
neously. If the soil meets only one of the criteria mentioned above,
further studies were suggested by Andrews and Martin [27].
Accordingly, even though the reconstituted silt utilized in this
study plots in the area defined as “susceptible to liquefaction”, the
point is almost on the borderline separating the “susceptible” and
Fig. 19. Effect of OCR on stiffness degradation of reconstituted silt. “not susceptible” sides, as shown in Fig. 21.
More recent criteria, based on the observations regarding the
sites where ground failures occurred in Adapazarı following the
1999 Kocaeli (Turkey) earthquake and accompanying laboratory
work were proposed by Bray et al. [28]. Consequently, the fine
grained soils were susceptible to liquefaction or cyclic mobility if
wc/LL Z0.85 and PI r12 are met; whereas the soils satisfying the
conditions of 0.8 rwc/LL r0.85 and 12 oPIr20 have been iden-
tified as moderately susceptible to liquefaction or cyclic mobility.
In a subsequent study, consisting of the investigations of field data
from the more recent earthquakes, Bray and Sancio [14] stated that
the clayey silts and silty clays having PI values between 12 and 18,
and wc/LL40.8 could undergo liquefaction, whereas sensitive soils
with PI 418 might undergo severe strength loss as a result of
Fig. 20. Chinese criteria for the soils having 0.005 mm and smaller particle sizes seismic loading. Plotting the relevant data for the reconstituted silt
less than 15%. utilized in this study so as to compare to the criteria they suggest
(Fig. 22), the silt is classified in the range from “moderately
reduction, however, is not as much significant as that for clays susceptible” to “susceptible” to liquefaction or cyclic mobility. This
shown by Vucetic and Dobry [23]. assessment is in conformity with the “cyclic mobility” observed for
the reconstituted silt during cyclic triaxial tests. The area defined
as “moderately susceptible” to liquefaction or cyclic mobility gives
5. Examination of the liquefaction susceptibility a room for evaluation of liquefaction susceptibility of borderline
of reconstituted silt via existing criteria materials, particularly for silts of low-plasticity. The criteria,
however, do not provide a clear distinction between the “liquefac-
There have been numerous definitions of soil liquefaction and tion” and “cyclic mobility”.
liquefaction susceptibility criteria, and unavoidably inconsistencies
among researchers over the years. The evaluation of liquefaction
susceptibility of fine grained soils began with the so called Chinese
Criteria introduced on the basis of the data collected following
large earthquakes in China [24]. Seed and Idriss [25] interpreted
Wang's findings and stated that clayey soils meeting the condi-
tions of (a) percent of particles less than 0.005 mm o 15%,
(b) liquid limit (LL)o 35, and (c) the ratio of initial water content
(wc) to the LL (wc/LL) 40.9 are likely to undergo severe strength
loss as a result of seismic loading. Koester [26] noted that LL values
determined by means of fall cone device used in China are about
four points higher than those values determined by means of
Casagrande cup used in U.S. Accordingly, LL value of 35 is taken
into consideration according to fall cone test applying the correc- Fig. 21. Liquefaction susceptibility criteria proposed by Andrews and Martin [27].
tions proposed by Koester [26] when considering Chinese lique-
faction assessment criteria.
The Chinese Criteria are represented in the form of a chart in
Fig. 20 for the soils having 0.005 mm and smaller particle sizes less
than 15%. The relevant data for the reconstituted silt used in this
study is as well plotted on the same figure. The silt, of which the
percent of particles less than 0.005 mm is about 9%, plots on the
“not susceptible to liquefaction” side of the chart. The assessment
of the criteria meets the test results when considering that no flow
liquefaction has been observed throughout the tests. However, this
assessment can be regarded as somewhat conservative when
considering the cyclic mobility type response that has frequently
occurred during the cyclic tests. It should also be noted that the
representative points are adjacent to the boundary defining
potential susceptibility to liquefaction. Fig. 22. Liquefaction susceptibility criteria proposed by Bray et al. [28].
174 S. Sağlam, B.Sadık Bakır / Soil Dynamics and Earthquake Engineering 61-62 (2014) 164–175
Boulanger and Idriss [1] recommended that the evaluation of
the liquefaction potential of fine-grained soils depended on the
behavior characteristically dominated by either clay (clay-like
behavior) or sand (sand-like behavior). Accordingly, while the
fine-grained soils with PI o7 have been classified as “sand-like”
(i.e. susceptible to liquefaction), soils with PI Z7 have been
classified as “clay-like”. Boulanger and Idriss [29] later stated that
if a soil plots on the plasticity chart as CL-ML the PI criterion may
be reduced to PIZ 5, and the soils with PI values of 3–6 are better
to be tested in-situ and in laboratory in addition to the liquefaction
correlations based on standard penetration test (SPT) and cone
penetration test (CPT). In accordance with this characterization,
the silt having PI ¼7 would be classified as “clay-like” material.
However, attention should be paid to the accuracy of the measured
parameters which becomes quite critical during liquefaction
susceptibility evaluation of such borderline materials.
6. Conclusions
An extensive laboratory testing program including conven-
tional soil mechanics tests, reconstitution procedure for soil
deposition, monotonic and cyclic triaxial testing was carried out
to evaluate the cyclic behavior of saturated low plastic silt. Based
on the results obtained through the laboratory tests, the key
findings are concluded as follows:
1. The cyclic response is observed to depend significantly on
whether the specimens are subjected to stress reversal or not
during cyclic loading. In the case of no stress reversal, plastic
strains accumulate with almost a constant rate in each cycle.
The plastic strain accumulation rate tends to decrease once the
peak cyclic stress becomes lower than the monotonic strength.
It is observed that the greater the ratio of the applied peak
cyclic stress to the monotonic strength, the greater the strain
accumulation rate. No significant cyclic degradation was
observed in stiffness of the specimens under loading without
stress reversals, although the axial strains exceeded 5%. In the
case of stress reversal, the strain accumulation is predominant
either in compression or in extension depending on the initial
stress state. The incremental strains developed in each addi-
tional cycle are added to the maximum past strains in com-
pression and extension respectively. The strain accumulation
rate becomes more pronounced in extension with increasing N,
which is attributed to lower strength of soils in extension.
2. Increasing confining stress has a critical effect on the cyclic
resistance of reconstituted NC and lightly OC silt.
3. While an initial shear stress up to 60–70% of p'i causes cyclic
strength to increase, cyclic strength was observed to decrease
with the initial shear stress beyond this range for the recon-
stituted silt specimens. It is observed that the cyclic resistance
increases more apparently for the tests with stress reversal, and
no significant influence of increasing initial shear stress is
detected for the case of no stress reversal. Additionally, the
maximum pore pressure ratio observed during shearing sig-
nificantly decreases with increasing τs/p'i.
4. The monotonic tests revealed that the undrained shear strength
increases with increasing OCR. Nevertheless, the increase in
resistance with increasing OCR observed during cyclic tests is
not as substantial as that observed in monotonic tests. On the
other hand, the maximum ru observed at each cycle decreases
with increasing OCR. If the cyclic resistance is evaluated by
means of pore water pressure generation, the cyclic resistance
increases more significantly with increasing OCR.
5. The loading rate effect differs depending on whether an initial
shear stress exists or not. An increase in the loading rate
obviously causes an increase in the cyclic resistance of silt
specimens that sustain no initial shear stress. In the general
sense, loading rate does not influence the cyclic resistance if
the cyclic behavior is dominated in the compression side. The
effect of loading frequency on the total pore pressure response
is significant for the isotropically consolidated specimens. A ten
folds increase in the loading frequency causes significant
decrease in the generation of excess pore pressure in the case
of isotropically consolidated specimens, whereas for the case of
no stress reversals its effect is not as significant as that
observed for the isotropically consolidated specimens.
6. The cyclic degradation increases with increasing CSRtx for all
OC levels applied during the tests. While an increase in OCR
reduces the stiffness degradation, this reduction is not as
significant as that for clays as shown earlier by Vucetic and
Dobry [23].
7. According to the Chinese Criteria, the silt utilized in this study
is evaluated as “not susceptible” to liquefaction. The assess-
ment of the criteria meets the test results when considering no
observation of flow liquefaction. However, this assessment can
be regarded as somewhat conservative when considering the
response of cyclic mobility that has frequently occurred during
the cyclic tests. The reconstituted silt is classified in the range
from “moderately susceptible” to “susceptible” to liquefaction
or cyclic mobility according to the criteria proposed by Bray
et al. [28]. This assessment is in conformity with the “cyclic
mobility” observed for the reconstituted silt during cyclic
triaxial tests. The area defined as “moderately susceptible” to
liquefaction or cyclic mobility provides a room for evaluation of
liquefaction susceptibility of borderline materials, especially for
low plasticity silts. However, the criteria do not provide a clear
distinction between the phenomena of “liquefaction” and “cyclic
mobility”.
References
[1] Boulanger RW, Idriss IM, Evaluating the potential for liquefaction or cyclic
failure of silts and clays, rep. no. UCD/CGM-04/01, Center of Geotechnical
Modeling, Dept. of Civil and Environmental Engineering, University of Cali-
fornia, Davis, California; 2004.
[2] Brandon TL, Rose AT, Duncan JM. Drained and undrained strength interpreta-
tion for low plasticity silts. J Geotech Geoenviron Eng 2006;132(2):250–7.
[3] Sanin MV, Wijewickreme D. Cyclic shear response of channel-fill Fraser River
Delta silt. Soil Dyn Earthquake Eng 2006;26:854–69.
[4] Mulilis JP, Seed HB, Chan CK, Mitchell JK, Arulanandan K. Effects of sample
preparation on sand liquefaction. J Geotech Eng Div 1977;103(GT2):91–108.
[5] Tatsuoka F, Ochi K, Fujii S, Okamoto M. Cyclic undrained triaxial and torsional
shear strength of sands for different sample preparation methods. Soils Found
1986;26(3):23–41.
[6] Yamamuro JA, Wood FM. Effect of depositional method on the undrained
behavior and microstructure of sand with silt. Soil Dyn Earthquake Eng
2004;24:751–60.
[7] Bishop AW, Henkel DJ. The measurement of soil properties in the triaxial test.
London, England: Edward Arnold Publishers LTD; 1957.
[8] Germaine JT, Ladd CC. Triaxial testing of saturated cohesive soils. Advanced
triaxial testing of soil and rock ASTM; 1988; 421–59.
[9] Andersen KH, Pool JH, Brown SF, Rosenbrand wF. Cyclic and static laboratory
tests on Drammen clay. J Geotech Eng Div, ASCE 1980;106(GT5):499–529.
[10] Yılmaz MT, Pekcan O, Bakır BS. Undrained cyclic shear and deformation
behavior of silt-clay mixtures of Adapazarı, Turkey. Soil Dyn Earthquake Eng
2004;24(7):497–507.
[11] Seed HB, Lee KL. Liquefaction of saturated sands during cyclic loading. J Soil
Mech Found Div 1966;92(6):105–34.
[12] Mitchell RJ, King RD. Cyclic loading of an Ottawa area Champlain Sea clay. Can
Geotech J 1977;14:52–63.
[13] Ishihara K. Soil behavior in earthquake geotechnics. Oxford: Clarendon Press;
1996.
[14] Bray JD, Sancio RB. Assessment of the liquefaction susceptibility of fine grained
soils. J Geotech Geoenviron Eng 2006;132(9):1165–77.
[15] Voznesensky EA, Nordal S. Dynamic instability of clays: an energy approach.
Soil Dyn Earthquake Eng 1999;18:125–33.
[16] Yasuhara K, Hirao K, Hyde AFL. Effects of cyclic loading on undrained strength
and compressibility of clay. Soils Found, J.G.S. 1992;32(1):100–16.
 S. Sağlam, B.Sadık Bakır / Soil Dynamics and Earthquake Engineering 61-62 (2014) 164–175
[17] Konrad JM, Wagg BT. Undrained cyclic loading of anisotropically consolidated
clayey silts. J Geotech Eng, A.S.C.E. 1993;119(5):929–47.
[18] Hyodo M, Yamamoto Y, Sugiyama M. Undrained cyclic shear behaviour of
normally consolidated clay subjected to initial static shear stress. Soils Found
Jpn Soc Soil Mech Found Eng 1994;34(4):1–11.
[19] Lefebvre G, LeBouef D. Rate effects and cyclic loading of sensitive clays. J
Geotech Eng, ASCE, 113; 1987; 476–89.
[20] Hyde AFL, Higuchi T, Yasuhara K. Liquefaction, cyclic mobility and failure of
silt. J Geotech Geoenviron Eng 2006;132(6):716–35.
[21] Polito CP, Green RA, Lee J. Pore pressure generation models for sands and silty
soils subjected to cyclic loading. J Geotech Geoenviron Eng, ASCE, 134; 2008;
1490–1500.
[22] Idriss IM, Dobry R, Singh RD. Non-linear behavior of soft clays during cyclic
loading. J Geotech Eng Div 1978;104(GT12):1427–47.
[23] Vucetic M, Dobry R. Degradation of marine clays under cyclic loading. J
Geotech Eng 1988;114(2):133–49.
175
[24] Wang W. Some findings in soil liquefaction. Beijing, China: Water Conser-
vancy, and Hydroelectric Power Scientific Research Institute; 1979.
[25] Seed HB, Idriss IM. Ground motions and soil liquefaction during earthquakes.
California, USA: Earthquake Engineering Research Institute, MNO-5 Berkeley;
1982.
[26] Koester JP. The influence of fines type and content on cyclic strength. . Ground
failures under seismic conditions, Geotechnical Special Publication, ASCE;
1994; 17–33.
[27] Andrews DCA, Martin GR, Criteria for liquefaction of silty soils. In: Proceedings
of 12th world conference on earthquake engineering, Auckland, New Zealand;
2000.
[28] Bray JD, Sancio RB, Durgunoğlu T, Önalp A, Youd TL, Stewart JP, et al.
Subsurface characterization at ground failure sites in Adapazarı, Turkey. J
Geotech Geoenviron Eng 2004;130(7):673–85.
[29] Boulanger RW, Idriss IM. Liquefaction susceptibility criteria for silts and clays. J
Geotech Geoenviron Eng 2006;132(11):1413–26.