Cyclic response of saturated silts

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: Received 15 December 2012 Received in revised form 16 August 2013 Accepted 19 February 2014

Softening and strength loss of sands with increasing excess pore water pressure under repeated loads is well-known. However, extensive damage to the built environment also occurs at the sites underlain by fine grained soils during seismic shaking. The primary objective of this study is to investigate the factors affecting cyclic behavior of saturated low-plastic silt through laboratory testing. For this purpose, an 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 loading characteristics on the cyclic response of saturated low-plastic reconstituted silt samples were examined separately. Based on the test results, a model was introduced to estimate the effect of initial shear stress on the cyclic response. Besides, liquefaction susceptibility of the samples was examined via current liquefaction susceptibility criteria. & 2014 Elsevier Ltd. All rights reserved.

Keywords: Cyclic triaxial test Liquefaction Silt

1. Introduction 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: [email protected], [email protected] (S. Sağlam). 1 Present address: Department of Civil Engineering, Adnan Menderes University, Aytepe, Aydın, Turkey. http://dx.doi.org/10.1016/j.soildyn.2014.02.011 0267-7261 & 2014 Elsevier Ltd. All rights reserved.

Hence, a detailed laboratory testing program has been undertaken 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 comparative 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 reconstitution procedure may have a substantial effect on the behavior of sands [4–6]. However, the information concerning the effect of reconstitution 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 monotonic 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 parameters, the specimens are tested under a range of preconsolidation 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.

S. Sağlam, B.Sadık Bakır / Soil Dynamics and Earthquake Engineering 61-62 (2014) 164–175

165

100 90

PERCENT FINER THAN D

80 70 60 50 40 30 20 10 0 0.001 CLAY

0.01

0.1

1

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 process The soil used in the tests was supplied in powdered form from Balad, Iraq. The grain size distribution of the light brown colored soil is presented in Fig. 1. The material consisted of 68.5% silt, 4.5% clay, and 27% fine sand size particles. The specific gravity is Gs ¼2.69, and Atterberg limits are determined as LL ¼31 (by means of Casagrande method), PL¼ 24 and PI¼ 7. The material is classified as low plasticity silt (ML) according to the Unified Soil Classification System (USCS), and plots adjacent to the A-line on the plasticity chart. During the process of reconstitution, the material was mixed with de-aired water of an amount required to bring the water content to about 2 to 3 times that of the liquid limit. Hence, the workability is improved during the mixing process, and consequently homogeneous slurry is provided. The slurry was then placed into a box having dimensions of 19.5, 19.5 and 21 cm, inside of which was covered with a woolen fabric and provided with narrow drainage holes on the top and bottom covers. The box was then submerged in de-aired water and the slurry was consolidated under a vertical pressure of 40 kPa, which was imposed by means of a pneumatic piston. Following the consolidation, 16 specimens could be extruded from the box and the applied consolidation pressure was sufficient to produce specimens that are able to 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.

3. Monotonic triaxial compression tests A series of undrained monotonic triaxial tests were conducted over the reconstituted samples to examine the response under monotonic loading, and to identify any conceivable relationship between monotonic and cyclic responses. An important issue during triaxial testing of fine grained soils is the rate at which the load is applied. Due to the friction at the two ends, distribution of stress and the strain is non-uniform along the specimen during loading. Sufficient time should be allowed during loading to provide stabilization of the pore water pressure within the specimen. Otherwise, the strength of the soil is affected by the nonuniformity of pore pressure. The monotonic triaxial tests were conducted at rates of strain ranging between 0.05 and 0.1%/min.

Table 1 Initial states of stress and loading rates of monotonic tests. Test

s'1c (kPa) s'3c (kPa) Initial p'i (kPa) Rate of εa (%/min) OCR ei

ST1 ST2 ST3 ST4 ST5 ST6 ST7 ST8 ST9 ST10 ST11 ST12 ST13 ST14 ST15 ST16 ST17 ST18 ST19 ST20 ST21 ST22 ST23 ST24 ST25

35 50 50 50 80 75 100 100 100 80 100 100 120 90 120 140 150 160 200 120 150 120 150 180 200

35 50 50 50 80 75 100 100 100 50 50 50 50 60 60 50 50 50 50 80 80 100 100 100 100

35 50 50 50 80 75 100 100 100 60 66.67 66.67 73.33 70 80 80 83.33 86.67 100 93.33 103.33 106.67 116.67 126.67 133.33

0.1 0.07 0.07 0.07 0.07 0.05 0.07 1 1.4 0.07 0.1 0.1 0.1 0.07 0.07 0.1 0.07 0.1 0.07 0.1 0.1 0.1 0.1 0.1 0.1

1 1 2 4 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

0.74 0.72 0.71 0.69 0.73 0.74 0.72 0.71 0.71 0.72 0.75 0.76 0.77 0.73 0.74 0.74 0.75 0.72 0.73 0.71 0.73 0.69 0.68 0.70 0.70

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]. Initial states of stress and applied loading rates for 25 strain controlled monotonic triaxial tests conducted on the reconstituted specimens are listed in Table 1. In addition to the tests performed with loading rates between 0.05 and 0.1%/min, two additional tests were carried out at a loading rate of 1.4 and 1%/min on the specimens isotropically consolidated under 100 kPa. Based on the test results, the average values of internal friction angle and cohesion were determined as 371 and 5 kPa, respectively. Two specimens with OCR of 2 and 4 were monotonically tested as well. State of overconsolidation was achieved through unloading the specimens that were isotropically consolidated under 100 and 200 kPa, respectively. Accordingly, both OC specimens sustained 50 kPa initial effective confining stress prior to shearing. 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 sustaining 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 monotonic shearing. The excess pore pressure generated during shearing 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 the influence of overconsolidation on the response to monotonic 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. 3. Monotonic pore pressure generation for the specimens with (a) no initial shear stress, and (b) a range of initial shear stresses.

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

Fig. 2. Monotonic stress paths for 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 generation decreases with increasing OCR and the pore pressure 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.

Fig. 5. Monotonic (a) stress–strain behavior and (b) pore pressure generation for the specimens with different OCR values.

dilate becomes more prevalent with increasing OCR during the monotonic response.

4. Cyclic triaxial tests Total of 69 cyclic tests were conducted over isotropically and anisotropically consolidated silt specimens. The consolidation pressures were intended to be representative of those likely to exist over the soil elements within the depths of interest beneath foundations. The range of cyclic stress ratios (CSR) were selected so as to commonly represent the amplitudes of loading during seismic events. Accordingly, a range of CSRtx values between 0.30 and 0.72 were utilized, and the corresponding amplitudes of cyclic stress were applied in a stresscontrolled manner. The ranges of stress and loading frequency; stress states previous to the cyclic phase; void ratio and the ratio of water content to LL at the beginning of cyclic shearing for the tested samples are presented in Table 2. Majority of the tests were conducted at a frequency of 0.5 Hz, which is considered as representative of the typical frequency range of seismically induced loading. In order to examine the influence of loading rate on the response, 11 of the tests were carried out at a frequency of 0.05 Hz. In order to investigate the influences of stress history and initial stress state on the cyclic response of silt, the tests were carried out with various values of OCR and initial shear stress ratios (τs/p'i). To impose the target OCR, following the initial isotropic consolidation phase under a specific confining stress, the drainage valve is closed and the confining pressure was appropriately reduced. Thus, the OC specimens had isotropic initial stress states ahead of being subjected to cyclic loading. On the other hand, the specimens in NC state were consolidated under various ratios of initial shear stress to initial mean effective stress (τs/p'i). The applied initial shear stress ratios over the specimens range between nil and 0.75. It should be noted that τs/p'i ¼0 condition refers to the isotropic

Test

ei s'1c (kPa) s'3c (kPa) τs/p'i CSRtx (Δscyc/2p'i)

w/LL

OCR

f (Hz)

C1 C2 C3 C4 C5 C6 C7 C8 C9 C10a C11 C12 C13a C14a C15 C16 C17 C18 C19 C20 C21 C22 C23 C24 C25a C26a C27 C28 C29 C30 C31 C32 C33 C34 C35 C36 C37 C38 C39 C40 C41 C42 C43 C44a C45 C46 C47a C48a C49a C50 C51 C52a C53 C54 C55 C56 C57 C58 C59 C60 C61 C62 C63 C64 C65 C66 C67 C68 C69

50 50 50 50 45 75 75 75 75 95 95 90 120 120 120 80 80 80 80 80 80 90 90 90 150 150 50 50 50 50 50 50 50 50 50 50 50 95 75 50 100 100 100 95 90 90 150 200 195 100 75 95 50 50 50 50 50 50 50 50 120 80 150 50 100 120 80 150 100

0.84 0.843 0.849 0.851 0.82 0.835 0.812 0.846 0.852 0.851 0.848 0.844 0.857 0.875 0.839 0.854 0.838 0.848 0.842 0.825 0.833 0.841 0.85 0.821 0.851 0.845 0.828 0.827 0.825 0.83 0.826 0.789 0.79 0.781 0.797 0.773 0.778 0.814 0.822 0.817 0.797 0.813 0.796 0.858 0.829 0.837 0.838 0.829 0.841 0.843 0.817 0.858 0.843 0.82 0.82 0.812 0.843 0.843 0.812 0.804 0.812 0.843 0.827 0.82 0.82 0.82 0.827 0.804 0.804

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 4 4 4 4 4 4 2 2.5 2 2 1.5 2 1 1 1 1 1 1 1 1 1 3 4 3 4 2 2 1 1 1 1 1 1 1 1 1 1 1

0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05

a

50 50 50 50 45 50 50 50 50 50 50 50 60 60 60 80 80 80 80 80 80 60 60 60 50 50 50 50 50 50 50 50 50 50 50 50 50 95 75 50 100 100 100 50 60 60 50 50 50 100 50 50 50 50 50 50 50 50 50 50 60 80 50 50 50 60 80 50 50

0 0 0 0 0 0.21 0.21 0.21 0.21 0.35 0.35 0.32 0.38 0.38 0.38 0 0 0 0 0 0 0.21 0.21 0.21 0.6 0.6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.35 0.21 0.21 0.6 0.75 0.74 0 0.21 0.35 0 0 0 0 0 0 0 0 0.38 0 0.6 0 0.38 0.38 0 0.6 0.38

0.35 0.3 0.55 0.6 0.72 0.34 0.51 0.69 0.69 0.35 0.54 0.71 0.41 0.5 0.59 0.31 0.41 0.5 0.59 0.59 0.59 0.36 0.54 0.64 0.45 0.6 0.3 0.5 0.6 0.6 0.65 0.35 0.35 0.5 0.55 0.6 0.8 0.45 0.47 0.55 0.55 0.48 0.3 0.4 0.48 0.61 0.7 0.39 0.6 0.34 0.4 0.42 0.5 0.62 0.7 0.53 0.52 0.58 0.53 0.85 0.54 0.54 0.6 0.3 0.37 0.44 0.44 0.48 0.56

0.74 0.74 0.75 0.75 0.72 0.74 0.72 0.75 0.75 0.75 0.75 0.74 0.75 0.77 0.74 0.75 0.74 0.75 0.74 0.73 0.73 0.74 0.75 0.72 0.75 0.74 0.73 0.73 0.73 0.73 0.73 0.7 0.7 0.69 0.7 0.68 0.69 0.72 0.72 0.72 0.7 0.72 0.7 0.76 0.73 0.74 0.74 0.73 0.74 0.74 0.72 0.76 0.74 0.72 0.72 0.72 0.74 0.74 0.72 0.71 0.72 0.74 0.73 0.72 0.72 0.72 0.73 0.71 0.71

Number of cycles was calculated considering SA axial strains.

consolidation. The sustained initial shear stresses were provided through increasing the deviatoric stresses (Δs) after consolidation 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 consolidated specimens subjected to two-way loading under different CSRtx and p'i values. The axial strain accumulation rate is observed to increase for CSRtx value of 0.31. There is no sudden increase in the 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 cycles of loading. In most of the isotropically consolidated cyclic tests, a loading rate of 0.5 Hz was used. The ru values, were, however, observed to reach, and even exceed 0.9 at different cycles. Although the excess pore pressure reaching at the initial confining stress led to 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. In those tests in which the anisotropically consolidated specimens subjected to increased loading demands (higher CSRtx values), the stress reversals also occurred. Fig. 8 shows the test results of anisotropically consolidated specimens subjected to stress reversals. The strains predominantly accumulate on the compression side of loading, and the accumulation rate increases with the increasing extent between the peak stress and the

Fig. 7. Stress–strain behavior of the specimens subjected to (a) CSRtx of 0.31 and p'i of 80 kPa (C16), (b) CSRtx of 0.50 and p'i of 80 kPa (C18).

monotonic strength as well as with the increasing CSRtx. The pore water pressure ratio ru generated in the specimen having an initial shear stress ratio τs/p'i of 0.21, approaches to 0.8, whereas it approaches to 0.5 for τs/p'i of about 0.35. However, majority of DA axial strains of the tests exceed 10% regardless of ru value.

4.1. Effect of initial confining stress In free field, soil elements are by and large subjected to uneven horizontal and vertical stresses. However, as indicated by Seed and Lee [11], the stress condition initially existing in free field is simulated through application of isotropic confining stress. Therefore, it is also of concern to distinguish the effects of initial shear 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. The effect of confining stress on the relationship between DA axial 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 show the relationship between CSRtx and N to reach 5% DA axial 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 with s'3c of 80 kPa. The effect of initial confining stress on cyclic response of NC reconstituted silt is observed to be parallel with those reported in the literature [12–14]. Nevertheless, this effect is examined for the lightly OC silt as well in the present study. The strain responses of

Fig. 10. Relationship between CSRtx and N to reach at 5% DA axial strain for reconstituted NC specimens consolidated under s'3c of 50 and 80 kPa.

Fig. 11. Relationship between DA axial strain and N for reconstituted specimens having OCR of 2 that were loaded with CSRtx of 0.55.

two OC specimens tested with CSRtx of 0.55 are illustrated in Fig. 11 as a function of N. The tests C40 and C41 with OCR of 2 were consolidated under s'3c of 50 and 100 kPa, respectively. The specimen tested with s'3c of 100 kPa undergoes high strains earlier than that tested with s'3c of 50 kPa. In Fig. 12, CSRtx is seen 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

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 The soil elements beneath existing foundations or in slopes are 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 previously 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 with reference to the sustained initial shear stresses. The cyclic 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 no stress reversal case.

The tests conducted with frequency of 0.5 Hz are considered 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 increasing 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 CSRtx ¼ c  N d

ð1Þ

Fig. 13. The relationship between CSRtx and N to reach at 5% εa for different τs/p'i values.

where the coefficients c and d are expressed in terms of

τs/p'i

Fig. 14. The relationship between (τs þτcyc)/p'i and N to reach at 5% εa for different τs/p'i values.

S. Sağlam, B.Sadık Bakır / Soil Dynamics and Earthquake Engineering 61-62 (2014) 164–175

171

Fig. 16. The relationship between CSRtx and N to reach at 5% εa for different OCR values.

Fig. 15. The model representing the effect of initial shear stress on cyclic response of reconstituted silt.

values as follows: c ¼ 0:7τs =p'i þ 0:5

ð2Þ

d ¼  0:18ðc þ 0:1Þ

ð3Þ

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.

Fig. 17. Excess pore pressure generation with axial strains for different OCR levels.

4.4. Effect of loading rate 4.3. Effect of overconsolidation ratio Stress history of the soils is an important factor influencing stress–strain response under both monotonic and dynamic loading. 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 monotonic 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.

Loading rate is another factor that has been frequently pronounced 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 generation during cyclic loading can be taken into consideration as separate components of transient and residual excess pore pressures [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

C3 C59 C5 C60 C18 C62 C17 C67 C10a C65a C13a C66a C11 C69 C15 C61 C26a C63a C25a C68a a

s'1c (kPa)

50 50 45 50 80 80 80 80 95 100 120 120 95 100 120 120 150 150 150 150

s'3c (kPa) τs/p'i

50 50 45 50 80 80 80 80 50 50 60 60 50 50 60 60 50 50 50 50

0 0 0 0 0 0 0 0 0.31 0.33 0.33 0.33 0.31 0.33 0.33 0.33 0.5 0.5 0.5 0.5

CSRtx

0.55 0.53 0.72 0.85 0.5 0.54 0.41 0.44 0.35 0.37 0.41 0.44 0.54 0.56 0.59 0.54 0.6 0.6 0.45 0.48

f (Hz)

0.5 0.05 0.5 0.05 0.5 0.05 0.5 0.05 0.5 0.05 0.5 0.05 0.5 0.05 0.5 0.05 0.5 0.05 0.5 0.05

At 5% εa ru (%)

N

68 51 73 35 25 55 51 40   71 76 48 46 30 58 41 27 55 52

7 1 25 1 2 1 6 1   46 91 3 3 2 2 9 2 113 39

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 specimens 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

Fig. 19. Effect of OCR on stiffness degradation of reconstituted silt.

Fig. 20. Chinese criteria for the soils having 0.005 mm and smaller particle sizes less than 15%.

reduction, however, is not as much significant as that for clays shown by Vucetic and Dobry [23].

5. Examination of the liquefaction susceptibility of reconstituted silt via existing criteria There have been numerous definitions of soil liquefaction and 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 conditions 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 corrections proposed by Koester [26] when considering Chinese liquefaction 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.

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 simultaneously. 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 “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 identified 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 seismic loading. Plotting the relevant data for the reconstituted silt utilized in this study so as to compare to the criteria they suggest (Fig. 22), the silt is classified in the range from “moderately susceptible” to “susceptible” to liquefaction or cyclic mobility. 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 gives a room for evaluation of liquefaction susceptibility of borderline materials, particularly for silts of low-plasticity. The criteria, however, do not provide a clear distinction between the “liquefaction” and “cyclic mobility”.

Fig. 21. Liquefaction susceptibility criteria proposed by Andrews and Martin [27].

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 conventional 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 additional cycle are added to the maximum past strains in compression 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 reconstituted 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 significantly 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 assessment 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 California, Davis, California; 2004. [2] Brandon TL, Rose AT, Duncan JM. Drained and undrained strength interpretation 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 Conservancy, 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.