Estimation of liquefaction potential from dry and saturated sandy soils under drained constant volume cyclic simple shear loading

Soil Dynamics and Earthquake Engineering 75 (2015) 27–36

Contents lists available at ScienceDirect

Soil Dynamics and Earthquake Engineering journal homepage: www.elsevier.com/locate/soildyn

Estimation of liquefaction potential from dry and saturated sandy soils under drained constant volume cyclic simple shear loading M.Murat Monkul a,n, Cihan Gültekin b,1, Müge Gülver c,1, Özge Akın d, Ece Eseller-Bayat d a

Department of Civil Engineering, Yeditepe University, İstanbul, Turkey 5A Mühendislik, İstanbul, Turkey c MAG Mühendislik, İstanbul, Turkey d Department of Civil Engineering, İstanbul Technical University, İstanbul, Turkey b

art ic l e i nf o

a b s t r a c t

Article history: Received 16 April 2014 Received in revised form 25 March 2015 Accepted 27 March 2015

Understanding the liquefaction mechanism of sandy soils still remains as one of the challenges in geotechnical earthquake engineering, since clean sands, silty sands and clayey sands do not necessarily show identical reactions under seismic loading. This study investigates the cyclic simple shear responses of three sandy soils: clean sand (Sile Sand 20/55), silty sand (Sile Sand 20/55 with 10% IZ silt) and clayey sand (Sile Sand 20/55 with 10% kaolin) based on many dry and saturated specimens. Drained constant volume cyclic simple shear tests on clean and silty sand specimens have shown that liquefaction potential of those soils could also be determined via dry samples. This is an important observation, since dry specimens are much easier to prepare and less time consuming compared to their saturated counterparts, as the demanding saturation process is eliminated. However, cyclic responses of dry and saturated clayey sand specimens were shown to be quite different, and therefore saturation of these specimens is still a must for liquefaction assessment. For both silt and kaolin, adding 10% fines to the base sand increased the liquefaction potential of resulting sandy soils considerably compared to the clean sand at the same void ratio. But this difference relatively decreased as the specimens became looser. & 2015 Elsevier Ltd. All rights reserved.

Keywords: Sand Clay Silt Simple shear testing Liquefaction Cyclic loading

1. Introduction Liquefaction of sandy soils is among the most investigated topics of modern geotechnical engineering but still remains to be one of the most challenging aspects of earthquake geotechnics. Field observations have shown that the predominant soil type in many of the liquefied sites were sandy soils involving certain amount of fines. Lade and Yamamuro [27] gave a summary of 20 cases of static or flow-type liquefaction and about 40 cases of seismic liquefaction. Those liquefaction cases occurred at various geotechnical structures including submarine slopes, mine tailings, hydraulic fills, spoil heaps, highway embankments, levees, level grounds and earth dams. Fourtynine out of fiftynine liquefaction case histories reported by Lade et al. [33] involved sands with certain amount of fines. Cetin et al. [13] analyzed 201 field case histories and proposed an SPT based probabilistic correlation for liquefaction assessment. One hundred and fiftyeight out of 201 case histories analyzed by Cetin et al. [13] involved sands with certain amount of fines (i.e. FC40).

n

Corresponding author. E-mail address: [email protected] (M.Mura. Monkul). 1 Formerly student at Department of Civil Engineering, Yeditepe University.

http://dx.doi.org/10.1016/j.soildyn.2015.03.019 0267-7261/& 2015 Elsevier Ltd. All rights reserved.

Bardet and Kapuskar [5] mentioned liquefaction of sandy soils at the Marina District of San Francisco during the 1989 Loma Prieta Earthquake. Stewart et al. [48] wrote about several liquefaction sites involving fines containing sands during the 1999 Chi-Chi Earthquake in Taiwan. Bray et al. [10] reported that severely damaged city of Adapazarı, Turkey during 1999 Kocaeli Earthquake has a soil profile generally involving loose silts and silty sands at shallow depths up to 5 m and some of the liquefied layers contain considerable amounts of clay sized particles as well. Bhattacharya et al. [8] expressed widespread liquefaction hazards due to sandy soils, which ranged from sandy silt to silty sand, at the Tokyo Bay area during the 2011 Tohoku Eartquake (the largest earthquake ever recorded in Japan, with Mw ¼9). Belkhatir et al. [7] denoted that significant loss of life and property occurred during 1980 Chlef Earthquake in Algeria, where various liquefaction cases involving sandy soils were observed. Taylor et al. [49] discussed the extensive damage at the Central Business District of Christchurch caused by liquefaction during the 2011 Christchurch Earthquake series. Accordingly, the upper 8 m of profile consists of soils ranging from silty fine sands to sandy silts with fines contents between 15% and 50%. Observations of liquefaction cases from various earthquakes mentioned above revealed two major aspects regarding the liquefaction research: 1) investigating the liquefaction of clean sands is important

28

M.Murat Monkul et al. / Soil Dynamics and Earthquake Engineering 75 (2015) 27–36

but not sufficient, as many of the in-situ liquefaction cases involve sandy soils involving certain silt or clay fractions. Hence, research should focus more on silty and/or clayey sands' behavior. 2) In parallel to the first aspect, laboratory investigation and liquefaction experiments in a controlled environment has still vital importance to understand the influence of various factors such as fines type, fines content, plasticity, grain size, shape and distribution etc. on liquefaction potential of sandy soils containing silt and/or clay as fine particles. To investigate the influence of such factors, it is often required to reconstitute sandy soil specimens with controlled parameters such as fines content, gradation characteristics, fines' plasticity etc. This brings together the various challenges of specimen preparation process such as convenience of the applied deposition technique for the soil of interest (e.g. segregation and distribution of fines within the specimen could be a problem in wet pluviation based techniques) and obtaining high and repeatable degrees of saturation with varying fines content and type. There have been substantial studies regarding the comparison of different depositional methods and corresponding undrained response of sands [42,54] and silty sands [21,60,11]. Meanwhile, obtaining high and comparable degrees of saturation for different laboratory deposited sandy soil specimens containing fines still remains to be one of the difficult and demanding tasks of reliable specimen preparation for liquefaction research. Obtaining fully saturated laboratory deposited specimens for liquefaction research also requires a very time consuming procedure. As an example, Monkul and Yamamuro [32] used the dry funnel deposition technique for deposition of triaxial silty sand specimens. After the deposition, specimens were flushed with CO2 in the dry state for 40 min and then de-aired water was percolated about 18.5 h to facilitate full saturation of specimens even before back pressure was applied. Hence, the saturation process of a single silty sand specimen took about 19 h for their study. Similarly, Hazirbaba [19] mentioned that silty sand specimens having more than 10% fines content (FC) required up to 48 h of back pressure application to achieve acceptable B values for cyclic simple shear testing. In 1977, Finn and Vaid [18] mentioned an important observation about the liquefaction behavior of Ottawa sand in constant volume cyclic simple shear condition. Accordingly, dry and saturated sand specimens showed identical drained constant volume simple shear behavior. This is an important statement; because the liquefaction potential of a sand could be predicted from dry specimens, eliminating the demanding and time consuming saturation process. In the study of Finn and Vaid [18], no experimental data or graph specifically comparing saturated specimens with their dry counterparts in constant volume simple shear were provided. Instead, it was clearly emphasized that no practical differences were found between the two. Up to the present, no experimental verification was available in the literature. Furthermore, it is unknown whether such a statement is valid for sands involving certain silt or clay fractions. The goal of this study is to investigate whether liquefaction potential of clean sands, silty sands and clayey sands could be predicted from dry soil specimens. Drained constant volume cyclic simple shear tests were performed on various fully saturated and dry sandy soil specimens. The resulting behaviors are compared and the possibility of predicting liquefaction potential from dry specimens is discussed depending on the soil type (i.e. clean sand, silty sand or clayey sand).

2. Cyclic simple shear testing and experimental setup Cyclic simple shear is probably the most popular laboratory test to investigate the dynamic behavior of soils after the cyclic triaxial testing. Cyclic simple shear testing has some advantages over cyclic triaxial testing. Cyclic loading mechanism in simple shear test resembles the earthquake loading conditions better compared to the cyclic

triaxial test. Consolidation in simple shear is anisotropic, and can be assumed to represent at rest condition in the field. Also specimen preparation is relatively easier compared to that for triaxial testing, mainly due to the smaller size of simple shear specimens with respect to the typical triaxial test specimens. Early developments of simple shear testing were initiated in UK [43] and Scandinavia at SGI [24] and at NGI [9]. Later, Peacock and Seed [38] adapted it to cyclic loading for liquefaction research. Two alternatives exist in modern simple shear testing to represent the volumetric conditions under dynamic loading in the field: “undrained” and “drained constant volume” testing. In undrained simple shear testing, the specimens could be enclosed in a pressure chamber similar to the triaxial testing. Some researchers kept the vertical stress constant [16], while some others kept the consolidated height of the specimen constant during shearing [20,14,22]. Regardless, the specimens are sheared in undrained conditions and the generated excess pore pressures are measured with a pore pressure transducer. In drained constant volume simple shearing, specimens are sheared in drained conditions in such a way that the volume of the specimens is kept constant during the entire shearing stage. Since, volumetric strain is equal to the axial strain in a simple shear test, constant volume is preserved by adjusting the magnitude of the vertical stress on the specimen so that the height of specimen does not change during shearing. Because drainage is allowed, no pore pressure is measured with a transducer, and the change of vertical stress is used to predict the excess pore pressures in an equivalent undrained test [9]. Dyvik et al. [16] demonstrated on normally consolidated clay that pore pressures measured in an undrained simple shear test is identical with the pore pressures predicted from a drained constant volume simple shear test. Following this verification, drained constant volume type of simple shear testing was also frequently used in liquefaction research [53,41,55,46,23]. In this study, an NGI type Geocomp cyclic simple shear device at Yeditepe University was used and the tests were done in drained constant volume condition. The typical specimens have a diameter of 64 mm and a height of 20 mm. The lateral confinement was provided by aligning several teflon coated rings around a conventional latex membrane. Typically, steel-wire-reinforced membranes are used in most NGI type devices in order to provide lateral confinement, even though ASTM [2], suggests to use either rings or a wire-reinforced membrane. Recently, Baxter et al. [6] compared the response of specimens confined with wire reinforced membrane and teflon coated rings with a Geocomp simple shear apparatus. Accordingly, teflon rings give more lateral stiffness during the consolidation stage, but the stress–strain response during shearing was similar for the two confinement systems. As mentioned before, pore pressure measurements and therefore the degree of saturation is not a concern for drained constant volume type of simple shearing (i.e. decrease in vertical stress to maintain constant height is equivalent to an increase in excess pore water pressure). Consequently, many of the typical simple shear devices of this type are not inherently designed for the challenging procedure of specimen saturation. Since the main goal of this study is comparing the responses of various fully saturated and dry sandy soil specimens, the bottom platen of the device in this study is modified in order to allow CO2 flushing and de-aired water percolation for obtaining fully saturated specimens. Note that the specimens in this study were either dry or fully saturated therefore; issues with partial saturation were not a concern during the discussion of the findings.

3. Soils tested and specimen preparation The base sand used in this study was obtained from a sand quarry at the Sile region of Istanbul and named as Sile Sand 20/55. This sand

M.Murat Monkul et al. / Soil Dynamics and Earthquake Engineering 75 (2015) 27–36

has a coefficient of uniformity (CU) of 2.79 and coefficient of curvature (CC) of 1.08 and classified as poorly graded sand (SP) according to the Unified Soil Classification System (USCS). Two different fine grained soils: a non-plastic silt and a low plastic clay were used in order to check the possible influence of different fine types on the test results. IZ silt is a naturally forming soil obtained from the city of Izmir, which has a natural fines content of 74%. Only the non-plastic –No 200 portion (o0.075 mm), obtained by wet sieving, was used in the experimental program. Kaolin was also used as an alternative fine grained soil, which has a liquid limit (LL) of 47.6% from the Casagrande method and a plasticity index (PI) of 10.9%. The kaolin used in this study falls in the region of low plastic silt (ML) in the plasticity chart. The grain size distributions of the three soils mentioned, obtained from sieve analysis and hydrometer tests, are given in Fig. 1a. The specific gravities were determined as 2.65, 2.70 and 2.59 for the Sile Sand 20/55, IZ silt, and Kaolin respectively. Sile Sand 20/55 was first thoroughly mixed with IZ silt on a dry weight basis to obtain silty sand with 10% fines content. Similarly, Sile Sand 20/55 was also thoroughly mixed with Kaolin on a dry weight basis to obtain clayey sand with 10% fines content. Grain size distribution curves for the resulting silty and clayey sands are shown in Fig. 1b. Influence of fines content is not among the main goals of this study. Therefore 10% FC is selected as a single but representative fraction, at which soil fabric is still dominated by the sand grain matrix. Maximum (emax) and minimum (emin) void ratios of the soils were determined by the method proposed by Lade et al. [28]. Maximum void ratio values were determined as 0.784 and 0.821 for the clean Sile Sand 20/55 and silty sand respectively. Whereas, minimum void ratio values were determined as 0.482 for the clean sand and 0.431 for the silty sand. Different methods of maximum and minimum void ratio determination (e.g. ASTM 2007 (D4253, D4254), Japanese standards etc.) could yield different emax and emin values, and also could involve their individual limitations (e.g. applicable up to a certain FC value). However, for this study the consistency and the repeatability of the procedure are important rather than the absolute values of emax and emin. Therefore, Lade et al. [28] method is employed considering that it had been successfully adopted in various previous research studies involving clean and silty sands [58,59,32,30]. The details of the method, which employs a calibrated graduated cylinder, are well explained by Lade et al. [28] and will not be repeated here. Before initiating the experimental program, several simple shear specimen preparation methods (i.e. wet pluviation, staged wet pluviation, slurry deposition and dry deposition) were compared based on various criteria including degree of saturation, distribution and achieved content of fines, repeatability and preparation time. The details

29

of this comparison and employed techniques can be found in Akın et al. [1]. Among those, the dry deposition based method was found to be most appropriate method for this study, because it enabled achieving greater degrees of saturation, closer values to the targeted fines content with less specimen preparation time and slightly greater repeatability. Akın et al. [1] calculated the degree of saturation of various dry deposited specimens based on water content and phase relationships. Accordingly, the average degree of saturation (S) for clayey sand specimens (with 10% Kaolin) obtained by this method was found to be 98% with a standard deviation (SD) of 1.2%. Also, the specimens were cut into four parts by slicing at mid height and at a cord through the center as shown in Fig. 2. Wet sieving was performed on each section with No 200 sieve to check whether fines are evenly distributed within the specimen. The distribution of fines over mentioned sections in Fig. 2 were found to be quite satisfactory and even (with an SD of 0.17%). The dry deposition technique used in this study is equivalent to the dry funnel deposition described in the literature for triaxial specimen preparation [26,62,31]. As shown in Fig. 3, a bottle shaped funnel with a curved tip was used. Once the funnel was placed on the filter paper above the porous stone, it was filled with dry soil. Then the funnel was raised gently along the axis of symmetry of the specimen allowing the soil to slowly fill into the space inside the rings with no drop height. Hence all specimens were deposited initially in a dry state. For saturated specimens, first CO2 was flushed from bottom to top for 15 min. After CO2 flushing, about 10 sample volumes of de-aired water was percolated through the specimens from bottom to top. De-aired water tank was located about 120 cm above the specimens, and de-aired water was percolated by the influence of gravity. Such a low pressure head and an extra head loss obtained with a control valve attached to the water entrance of the specimen ensured the low gradient to minimize the migration of fines grains during the saturation process. In fact, this was experimentally verified with wet sieving as explained in Fig. 2, where the distribution of fines over the top and bottom sections of the specimen was found to be even. After the deposition process was completed, all specimens were consolidated to the initial vertical effective stress (σ0 vc) of 50 kPa. During the cyclic loading stage, uniform sinusoidal cycles of shear stresses (τcy) were applied at a cyclic stress ratio (CSR) of 0.12 (i.e. τcy/σ0 vc ¼ 0.12) with a frequency of 0.1 Hz. No monitoring or corrections were done for the influence of the filter paper (e.g. compressibility) in this study. However, considering that the initial vertical effective stress (σvc ¼50 kPa), CSR (0.12), loading frequency (0.1 Hz), testing apparatus, and filter paper type were kept the same for all the specimens tested,

Fig. 1. (a) Grain size distribution of Sile Sand 20/55, IZ silt and Kaolin used in the experimental program. (b) Grain size distribution curves for clean sand, silty sand (with 10% IZ silt) and clayey sand (with 10% kaolin).

30

M.Murat Monkul et al. / Soil Dynamics and Earthquake Engineering 75 (2015) 27–36

Fig. 2. Method employed by Akin et al. [1] to investigate the distribution of fines in simple shear specimens.

Fig. 4. Change of number of cycles to liquefaction with consolidated void ratio for clean Sile Sand 20/55 (σ0 vc ¼ 50 kPa, CSR ¼0.12).

Fig. 3. Specimen preparation with dry funnel deposition technique in this study.

the possible influence of filter paper would not alter the observed trends and comparisons in the findings. A total of 53 drained constant volume cyclic simple shear tests were conducted in this study. Liquefaction was considered to occur when the predicted excess pore pressure was equal to the initial vertical effective stress (i.e. Δue ¼ σvc). Tests were continued until either the specimens have liquefied or 10% double amplitude strain was reached, whichever occured first.

4. Experimental results and discussion 4.1. Cyclic simple shear tests on clean Sile Sand 20/55 Drained constant volume cyclic simple shear tests were initially carried out on saturated clean Sile Sand 20/55 specimens. Tests were done at various void ratios within a range between 0.6 and 0.72. Neither of the tamping, tapping with a tool, or vibration was employed to achieve a wider range of densities, as those densification techniques could influence the initial fabric and therefore the resulting shearing response of the specimens, especially for sands with fines [31]. Instead, different void ratios were obtained by changing the funnel raising speed. Generally, slightly denser specimens were obtained with increasing funnel raising speed, but in all cases no drop height between the tip of the funnel and the deposited soil was allowed. Once the tests with saturated specimens were completed, the tests on dry specimens were done. The consolidated void ratios versus number of cycles to liquefaction for both dry and saturated specimens of Sile Sand 20/55 are shown in Fig. 4. The number of cycles to liquefaction increases with decreasing void ratio (liquefaction resistance increases with increasing density). Fig. 4 also reveals that liquefaction behavior of both dry and saturated clean sand specimens are fairly close to each other under drained constant volume cyclic simple shear loading and could be

represented by a single trend curve. This is in agreement with the statement of Finn and Vaid [18] that, no practical difference exist between the two. Fig. 5 presents the cyclic response of the two loose sand specimens circled in Fig. 4. The dry sand specimen had a consolidated void ratio of 0.722 and a relative density (Dr) of 20.6% after consolidation. The saturated sand specimen had a consolidated void ratio of 0.72 and a Dr of 21.3% after consolidation. So, the specimens shown in Fig. 5 represent the loose sand behavior under drained constant volume cyclic simple shearing. The trends of hysteresis loops of both specimens in Fig. 5a are similar, even though in the third cycle positive shear strains were more pronounced for the dry specimen, while negative shear strains were more pronounced for the saturated one. Both specimens have shown considerable strain softening after the second cycle. The general patterns of cyclic stress paths of both specimens shown in Fig. 5b are also close to each other, even though some deviations are observed in the second cycle. The progress of shear strains with cycles is shown in Fig. 5c, where negative shear strains were relatively pronounced starting from the end of the second cycle for the saturated specimen. The generation of predicted excess pore pressures with cyclic loading for both specimens shows a very good agreement (Fig. 5d) and both specimens are considered as liquefied in the third cycle. The specimens compared in Fig. 5 were loose and therefore had a low number of cycles to liquefaction (i.e. N¼3). One can wonder about the response of denser specimens where the number of cycles to liquefaction is greater. For this purpose, the cyclic responses of two medium dense specimens circled in Fig. 4 were compared in Fig. 6. It should be noted that these two specimens do not have exactly the same void ratio (or relative density) or the same number of cycles to liquefaction, but they are reasonably close to each other as seen in Fig. 4. The dry sand specimen had a void ratio of 0.61 and a relative density of 57.6% after consolidation and had the number of cycles to liquefaction of 16. The saturated sand specimen had a void ratio of 0.599 and a Dr of 61.4% after consolidation and had the number of cycles to liquefaction of 15. So, the specimens shown in Fig. 6 represent the medium dense sand behavior under drained constant volume cyclic simple shearing. The trends of hysteresis loops of both specimens in Fig. 6a are similar, where more strain softening was observed for the last four cycles for both specimens. The general patterns of cyclic stress paths of both specimens shown in Fig. 6b are also reasonably close to each other, even though some deviations exist. The progresses of shear strains with cycles shown in Fig. 6c are in good agreement, where considerable shear strains started to accumulate after the tenth cycle for both specimens. The generation of predicted excess pore pressures with cyclic loading for both specimens is shown in Fig. 6d. A temporary drop in excess pore

M.Murat Monkul et al. / Soil Dynamics and Earthquake Engineering 75 (2015) 27–36

31

Fig. 5. (a). Cyclic stress–strain response of dry (e¼ 0.722, Dr ¼ 20.6%) and saturated (e ¼0.72, Dr ¼ 21.3%) clean sand specimens (σ0 vc ¼50 kPa, CSR¼ 0.12). (b) Cyclic stress paths of dry (e¼0.722, Dr ¼ 20.6%) and saturated (e ¼0.72, Dr ¼21.3%) clean sand specimens (σ0 vc ¼50 kPa, CSR ¼0.12). (c) Shear strain versus number of cycles to liquefaction for dry (e¼ 0.722, Dr ¼20.6%) and saturated (e¼ 0.72, Dr ¼ 21.3%) clean sand specimens (σ0 vc ¼ 50 kPa, CSR ¼0.12). (d) Predicted excess pore water pressures versus number of cycles to liquefaction for dry (e¼ 0.722, Dr ¼20.6%) and saturated (e¼ 0.72, Dr ¼21.3%) clean sand specimens (σ0 vc ¼ 50 kPa, CSR¼ 0.12).

pressures for both specimens was observed within the first couple of cycles and there is a good agreement between the predicted excess pore pressures of the two specimens in the entire cyclic stage. Consequently, as discussed based on Figs. 4–6, the practical difference between the liquefaction behavior of dry and saturated clean Sile Sand 20/55 under drained constant volume simple shear loading can be considered as negligible.

4.2. Cyclic simple shear tests on Sile Sand 20/55 with 10% IZ silt Tests were also conducted on dry and saturated silty sand specimens obtained by mixing Sile Sand 20/55 with 10% IZ silt. Specimens were tested at various consolidated void ratios within a range between 0.59 and 0.74. The change of number of cycles to liquefaction with void ratio is shown in Fig. 7. In the same figure the clean sand curve obtained from Fig. 4 is also plotted for comparison. Similar to the clean sand, the liquefaction resistance of silty sand specimens increases with decreasing void ratio. But the trend curve for silty sand is much flatter than the trend curve for clean sand. This means that for the same amount decrease in void ratio, the increase in liquefaction resistance is much smaller for the silty sand than for the clean sand. It is also important to note that, adding 10% silt to the sand considerably increased the liquefaction potential compared to the clean sand at the same void ratio, as the trend curve is shifted downwards for the silty sand. However, the difference between the liquefaction potentials of clean and silty sands decreases, as the specimens became looser (Fig. 7).

Fig. 7 also shows that liquefaction potentials obtained for both dry and saturated silty sand specimens are fairly close to each other under drained constant volume cyclic simple shear loading and they could be represented by a single trend curve. In order to inspect the results closer, the cyclic response of the two silty sand specimens, circled in Fig. 7, were presented in Fig. 8. Both the dry and saturated silty sand specimens had void ratios of 0.666 and relative densities (Dr) of 39.8% after consolidation, which represent loose to medium dense silty sand behavior under drained constant volume cyclic simple shearing. The trends of hysteresis loops of both specimens in Fig. 8a are similar. However, the saturated specimen has shown more strain softening after the second cycle and liquefied in the third cycle. The general patterns of cyclic stress paths of both specimens shown in Fig. 8b are also close to each other, but the saturated specimen liquefied a cycle before the dry specimen. The progress of shear strains with cycles is shown in Fig. 8c, where the responses of both specimens were quite close especially for the first two cycles. The generation of predicted excess pore pressures with cyclic loading for both specimens shown in Fig. 8d is also reasonably close to each other. Consequently, as discussed based on Figs. 7 and 8, the practical difference between the liquefaction behavior of dry and saturated silty sand specimens under drained constant volume simple shear loading can be considered as negligible. It should be noted that, neither in clean sand, nor in silty sand, the stress–strain responses of saturated specimens are perfectly identical with the responses of dry specimens (Figs. 5, 6 and 8). Yet, this is thought to be normal considering that some scatter is also inherently available within saturated specimens (Figs. 4 and 7), as well as within the dry specimens themselves (i.e. the

32

M.Murat Monkul et al. / Soil Dynamics and Earthquake Engineering 75 (2015) 27–36

Fig. 6. (a) Cyclic stress–strain response of dry (e¼ 0.61, Dr ¼ 57.6%) and saturated (e¼ 0.599, Dr ¼ 61.4%) clean sand specimens (σ0 vc ¼ 50 kPa, CSR ¼0.12). (b). Cyclic stress paths of dry (e ¼0.61, Dr ¼57.6%) and saturated (e¼ 0.599, Dr ¼ 61.4%) clean sand specimens (σ0 vc ¼50 kPa, CSR ¼0.12). (c) Shear strain versus number of cycles to liquefaction for dry (e¼0.61, Dr ¼57.6%) and saturated (e¼0.599, Dr ¼ 61.4%) clean sand specimens (σ0 vc ¼ 50 kPa, CSR ¼0.12). (d) Predicted excess pore water pressures versus number of cycles to liquefaction for dry (e¼ 0.61, Dr ¼57.6%) and saturated (e ¼0.599, Dr ¼61.4%) clean sand specimens (σ0 vc ¼ 50 kPa, CSR ¼0.12).

Fig. 7. Change of number of cycles to liquefaction with consolidated void ratio for Sile Sand 20/55 mixed with 10% IZ silt (σ0 vc ¼50 kPa, CSR¼ 0.12). (For interpretation of the references to color in this figure, the reader is referred to the web version of this article.)

responses of two saturated specimens at the same void ratio might not be perfectly identical either). On the other hand, Fig. 4 for clean sand and Fig. 7 for silty sand clearly demonstrate that, both dry and saturated specimens can be represented by the same trend curve specific for each soil type. Hence, dry specimens of clean sands and silty sands could be used to predict the liquefaction potential of those soils by drained constant volume cyclic simple shear tests, without the necessity of demanding

saturation effort and time. This conclusion is also consistent with a limited number of previous observations in literature. Wijewickreme et al. [56] performed a control test on a dry and a saturated air pluviated specimen of Fraser River sand under drained constant volume cyclic simple shear and observed that the responses were similar. Based on this observation, they continued their experimental program with dry air pluviated specimens to use in numerical modeling of centrifuge tests. Monkul et al. [33] conducted drained, stress path controlled triaxial tests on several dry and saturated specimens of a silty sand (Nevada Sand with 20% Loch Raven silt), in order to represent stress conditions in slopes due to changing groundwater conditions. Initially, the specimens were sheared in conventional drained triaxial compression stress path until the desired deviator stress (q) was reached, and then a constant shear stress path (constant q with decreasing effective mean normal stress (p0 )) was followed until the failure. They observed that the response between dry and saturated silty sand specimens were similar. 4.3. Cyclic simple shear tests on Sile Sand 20/55 with 10% kaolin In the final stage of the experimental program, tests were performed on dry and saturated clayey sand specimens obtained by mixing Sile Sand 20/55 with 10% kaolin. Specimens were tested at various consolidated void ratios. The change of number of cycles for liquefaction with void ratio is shown in Fig. 9. It is very clear from Fig. 9 that, unlike clean sand and silty sand specimens, the cyclic response of saturated and dry clayey sand specimens are distinctly different. It is seen in the figure that, the saturation process decreased the void ratio of specimens significantly compared to the dry ones,

M.Murat Monkul et al. / Soil Dynamics and Earthquake Engineering 75 (2015) 27–36

33

Fig. 8. (a) Cyclic stress–strain response of dry and saturated silty sand specimens (e ¼0.666, Dr ¼ 39.8%, σ0 vc ¼50 kPa, CSR ¼0.12). (b) Cyclic stress paths of dry and saturated silty sand specimens (e ¼0.666, Dr ¼ 39.8%, σ0 vc ¼ 50 kPa, CSR¼ 0.12). (c) Shear strain versus number of cycles to liquefaction for dry and saturated silty sand specimens (e¼ 0.666, Dr ¼ 39.8%, σ0 vc ¼ 50 kPa, CSR¼ 0.12). (d) Predicted excess pore water pressures versus number of cycles to liquefaction for dry and saturated silty sand specimens (e¼ 0.666, Dr ¼ 39.8%, σ0 vc ¼ 50 kPa, CSR ¼0.12).

Fig. 9. Change of number of cycles to liquefaction with consolidated void ratio for Sile Sand 20/55 mixed with 10% kaolin (σ0 vc ¼ 50 kPa, CSR ¼0.12).

considering that they were both initially consolidated to the same vertical effective stress of 50 kPa. It was observed that the total volume of specimens in Fig. 9 have decreased during the saturation process. Such a decrease in specimen volume during saturation had also been reported by several researchers investigating the behavior of sands with fines [47,57,32] which could be attributed to the elimination of some of the metastable grain contacts located between the sand grain matrix. The reason of decreasing void ratio for saturated specimens in Fig. 9 can then be explained with the phase relationships. As an example, the volume of solids (Vs) are the same and does not change

for the two identically prepared dry funnel deposited specimens, but the total volume decreases for the saturated specimen during the saturation process as explained above. Therefore, the volume of voids (Vv) decreases, while solid volume (Vs) remains the same during the saturation process. This had caused a corresponding decrease in void ratio (i.e. e¼ Vv/Vs) of the saturated specimens compared to their identically deposited dry counterparts in Fig. 9. When Fig. 7 is investigated carefully, one could observe a similar but less visible trend for the saturated silty sand specimens shown with yellow dots (e.g. no saturated specimens were achieved for void ratios around 0.7 and greater). One can initially expect an increase in total volume of the specimens in Fig. 9 during saturation, because the clay matrix within the sand would expand as it adsorbs water. Even though such a mechanism occurs, because of the small amount of fines content (10%) and low plasticity of the kaolin (PI¼10.9%) used in this study, the net change in total volume was dominated by the contraction during saturation as explained before, rather than the expansion of the clay matrix within the specimens shown in Fig. 9. Even though such a noticeable decrease in void ratio occurs, saturated specimens had considerably greater liquefaction potential compared to the dry specimens (Fig. 9). As an example, when the trend curve for dry clayey sand specimens in Fig. 9 is extrapolated towards the void ratio of 0.6, the approximate number of cycles to liquefaction is read as 22, while it is only 3 for a saturated specimen at the same void ratio. The exact explanation for such a dramatic increase in liquefaction potential of clayey sand specimens with saturation is not known. But it could be related to the initial fabrics achieved before cyclic loading. As the water molecules are attracted to the negatively charged surfaces of the clay particles, a lubrication effect could have

34

M.Murat Monkul et al. / Soil Dynamics and Earthquake Engineering 75 (2015) 27–36

occurred, which would decrease the frictional resistance compared to specimens with dry clay particles. The same mechanism would not occur in clean or silty sands, as the silt (non-plastic in this study) and sand grains are chemically inert. It should be noted that even though the kaolin has a PI of 10.9% as mentioned before, the resulting clayey sand with 10% kaolin is still non-plastic as 90% of the soil is composed of sand grain matrix. The trend curve for saturated clayey sand is flatter than the trend curve for dry clayey sand (Fig. 9); meaning that for the same amount of decrease in void ratio, the increase in liquefaction resistance is greater for the dry specimens than for the saturated ones. In Fig. 9, the clean sand curve obtained from Fig. 4 is also plotted for comparison. For saturated specimens, adding 10% kaolin to the sand significantly increased the liquefaction potential compared to the clean sand at the same void ratio, as the trend curve is shifted markedly downwards for the saturated clayey sand specimens. However, this difference between the liquefaction potentials of the saturated specimens of clean and clayey sands decreases, as the specimens became looser based on the orientations of trend curves in Fig. 9. Interestingly, adding 10% kaolin to the sand increased the resistance of dry specimens compared to the clean sand at the same void ratio, as the trend curve is shifted upwards. And this difference in resistance apparently increases as the specimens become looser. Whether similar trends would be observed for greater clay contents is currently unknown, as the initial fabric of specimens could be more strongly influenced by the content and plasticity characteristics of the clay matrix at greater fines contents such as 20% or 30%. But this is beyond the scope of the current investigation and could be a good direction for future studies. Various parameters such as void ratio, relative density (Dr), intergranular void ratio (void ratio of the sand matrix, es) were used in previous research in order to compare the liquefaction behaviors of clean sands versus sands with fines (e.g. the influence of FC). Perhaps, the most commonly used parameter is the void ratio for such a comparison [52,39,17,37,45,20], since it is one of the key parameters in geotechnical engineering. Some researchers used loosest possible density after deposition [61,26,35,3,32,34] as a comparison basis. In this method, liquefaction behaviors of various sandy soils are not necessarily compared at the same void ratio, instead those soils are assumed to be deposited at a “quasi natural void ratio”, provided that the depositional method is kept the same. Alternatively, some researchers used relative density [12,20] and some reported that Dr is the most appropriate parameter for such a comparison [40]. Some researchers employed intergranular void ratio, which is the void ratio of the sand matrix alone when fines are considered as voids [44,25,50,20,15]. The parameter of intergranular void ratio is shown in Eq. 1, where e is the void ratio, G is the specific gravity of the overall soil (weighted average of sand and fine constituents can be used), Gf is the specific gravity of the fines, FC is the percent of fines content by dry weight.

es ¼

FC e þ GGf :100 FC 1  GGf :100

ð1Þ

One might expect that employing intergranular void ratio could be better than void ratio since the sand matrix is dominant for the soils tested in this study (i.e. FC¼10%). However, Monkul and Yamamuro [32] had experimentally shown for different silty sands that employing intergranular void ratio is not necessarily a better parameter for comparing the liquefaction potential of sandy soils (i.e., an increase in intergranular void ratio does not necessarily imply an increase in liquefaction potential of the soil). This is due to the altered soil fabric that could be influenced by both the content and type of the fines, even though the base sand is kept the same. The change of number of cycles to liquefaction with intergranular void ratio values calculated by

Fig. 10. Change of number of cycles to liquefaction with consolidated intergranular void ratio for various sandy soils tested (σ0 vc ¼50 kPa, CSR ¼ 0.12).

Eq. 1 is shown in Fig. 10 for clean, silty and clayey sands tested in this study. Fig. 10 shows that for a specific soil type, number of cycles to liquefaction decreases with increasing intergranular void ratio (i.e. liquefaction potential increases with increasing es). However, such a trend is not valid when liquefaction behaviors of different sandy soils in Fig. 10 are compared at the same es, even though base sand and fines content is the same. For example, with a small extrapolation of trend curves, liquefaction potentials of clean, silty and clayey sands can be compared at es ¼0.8. Clean sand and clayey sand specimens are observed to liquefy at similar number of cycles, however silty sand specimen showed greater resistance to liquefaction at es ¼0.8. Indeed, several other researchers made significant contributions to the concept of intergranular void ratio and proposed to employ a modified version of it, where fines are not entirely considered as voids and the influence of fine grain matrix is also adjusted with an extra parameter added to Eq. 1 [51,36,4]. All of the mentioned literature, which employed void ratio, relative density or intergranular void ratio as a comparison basis for the liquefaction behavior of sandy soils, is valuable and legitimate. This study does not attempt to assess which comparison parameter is better; therefore void ratio was selected and used, as it is probably the most commonly used density parameter in both geotechnical research and practice.

5. Summary and conclusions In this study, cyclic simple shear responses of three sandy soils: clean sand (Sile Sand 20/55), silty sand (Sile Sand 20/55 with 10% IZ silt) and clayey sand (Sile Sand 20/55 with 10% kaolin) were investigated in order to compare the responses of dry and saturated specimens. A total of 53 stress controlled, drained constant volume cyclic simple shear tests were conducted on various dry and saturated specimens. Experiments on clean sand specimens demonstrated that liquefaction resistance of both dry and saturated sand specimens can be represented by the same trend curve in a number of cycles versus void ratio diagram. Cyclic responses of loose and medium dense saturated sand specimens were compared with their dry counterparts in terms of stress–strain, cyclic stress paths, and the development of shear strains and excess pore pressures with number of cycles, and the difference is found to be practically negligible. Similar to the clean sand response, experiments on silty sand specimens demonstrated that liquefaction resistance of both dry and saturated silty sands can be represented by the same trend curve in a number of cycles versus void ratio diagram. The cyclic response of a medium to dense saturated silty sand specimen was compared with a dry specimen in terms of stress–strain, cyclic stress paths, and the

M.Murat Monkul et al. / Soil Dynamics and Earthquake Engineering 75 (2015) 27–36

development of shear strains and excess pore pressures with number of cycles, and the difference is found to be practically negligible as well. It is also observed that, liquefaction potentials of silty sand specimens (with 10% FC) were considerably greater than the clean sand specimens at the same void ratio. But, this difference decreases as the specimens became looser. However, tests on clayey sand specimens do not obey the conclusion found for clean and silty sands, as cyclic response of dry and saturated clayey specimens were quite different (they cannot be represented by the same trend curve in a number of cycles versus void ratio diagram). It is observed that saturated clayey sand specimens had considerably greater liquefaction potential compared to the dry ones. For saturated specimens, it is observed that, liquefaction potentials of clayey sand specimens (with 10% FC) were considerably greater than the clean sand specimens at the same void ratio. But, this difference decreases as the specimens became looser. The above conclusions imply that the liquefaction potentials of clean and silty sands can be determined from dry specimens with drained constant volume simple shear tests. This is an important finding as: 1) preparing dry specimens would be much easier compared to the saturated ones as the demanding effort for saturation process would be eliminated, 2) possible uncertainties and waste of specimens due to differences in degree of saturations could be prevented, 3) specimen preparation durations could significantly be shortened, as no water de-airation, CO2 flushing and several sample volumes of de-aired water percolation is needed. For clayey sands, unfortunately and perhaps somewhat expectedly, saturation of specimens is still a must for liquefaction assessment. Some limitations of the above findings should also be mentioned. The conclusions in this study are based on cyclic behavior of sandy soils at low fines contents (FCr10%). Tests and verifications at greater fines contents could be a good direction for future research, as the initial fabric and its sensitivity to saturation (because of the interaction with water) could be influenced by fines content. Also, one should be careful about the consistency of the specimen preparation method. Both dry and saturated specimens in this study were deposited with the very same dry deposition technique. It is well known that, different deposition methods could constitute saturated specimens of the same soil with very different shear responses because of the different initial fabrics achieved [60,29]. As an example; cyclic response of dry specimens deposited with the dry funnel deposition method could be different from their saturated counterparts deposited with slurry deposition technique, as those deposition techniques could yield different initial fabrics. On the other hand, the results in this study imply that the behavior of a dry clean or silty sand specimen in drained constant volume cyclic simple shear testing can expected to be reflective for the behavior of an equivalent saturated centrifuge specimen prepared with the same initial deposition method such as dry deposition.

[7]

[8]

[9] [10]

[11] [12]

[13]

[14] [15] [16] [17] [18]

[19]

[20]

[21]

[22]

[23] [24] [25]

[26] [27] [28] [29]

[30] [31]

References [1] Akın, O, Monkul, MM, Eseller-Bayat, E. “Homogenous and saturated specimen preparation methods for simple shear test”. In: Proceedings of the 5th geotechnical symposium. Paper no:27, Cukurova University, Adana, Türkiye (CD-ROM) (In Turkish); 2013. [2] ASTM (American Society for Testing and Materials). Standard test method for consolidated undrained direct simple shear testing of cohesive soils. Philadelphia, USA: ASTM; 2007 D6528-07. [3] Bahadori H, Ghalandarzadeh A, Towhata I. Effect of non plastic silt on the anisotropic behavior of sand. Soils FoundJGS 2008;48(4):531–45. [4] Baki Md A,L, Rahman MM, Lo SR, Gnanendran CT. Linkage between static and cyclic liquefaction of loose sand with a range of fines content. Can Geotech J 2012;49:891–906. [5] Bardet JP, Kapuskar M. Liquefaction sand boils in San Francisco during 1989 Loma Prieta Earthquake. J Geotech Eng 1993;119(3):543–62. [6] Baxter CDP, Bradshaw AS, Ochoa-Lavergne M, Hankour R. DSS test results using wire-reinforced membranes and stacked rings. In: Proceedings of the

[32] [33] [34] [35] [36] [37]

[38]

35

advances in analysis modeling and design, ASCE. Geotechnical Special Publication 199; 2010. p. 600–7. Belkhatir M, Arab A, Della N, Schanz T. Experimental study of undrained shear strength of silty sand: effect of fines and gradation. Geotech Geol Eng 2012; 30(5):1103–18. Bhattacharya S, Hyodo M, Goda K, Tazoh T, Taylor CA. Liquefaction of soil in the Tokyo Bay area from the 2011 Tohoku (Japan) Earthquake. Soil Dyn Earthq Eng 2011;31:1618–28. Bjerrum L, Landva A. Direct simple shear tests on a Norwegian quick clay. Geotechnique 1966;16(1):1–20. Bray JD, Sancio RB, Durgunoglu T, Onalp A, Youd TL, Stewart JP, et al. Subsurface characterization at ground failure sites in Adapazari, Turkey. J Geotech Geoenviron Eng ASCE 2004;130(7):673–85. Carraro JAH, Prezzi M. A new slurry-based method of preparation of specimens of sand containing fines. Geotech Test J 2008;31(1):1–11. Carraro JAH, Prezzi M, Salgado R. Shear strength and stiffness of sands containing plastic or nonplastic fines. J Geotech Geoenviron Eng ASCE 2009;135(9):1167–78. Cetin KO, Seed RB, Kiureghian A, Tokimatsu K, Harder Jr. LF, Kayen RE, et al. Standard penetration test-based probabilistic and deterministic assessment of seismic soil liquefaction potential. J Geotech Geoenviron Eng ASCE 2004;130 (12):1314–40. Chang WJ, Hong ML. Effects of clay content on liquefaction characteristics of gap-graded clayey sands. Soils Found 2008;48(1):101–14. Dash HK, Sitharam TG, Baudet BA. Influence of non-plastic fines on the response of a silty sand to cyclic loading. Soils Found 2010;50(5):695–704. Dyvik R, Berre T, Lacasse S, Raadim B. Comparison of truly undrained and constant volume direct simple shear tests. Géotechnique 1987;37(1):3–10. Erten D, Maher MH. Cyclic undrained behavior of silty sand. Soil Dyn Earthq Eng 1995;14:115–23. Finn, WDL, Vaid, YP. Liquefaction potential from drained constant volume cyclic simple shear tests. In: Proceedings of the sixth world conference on earthquake engineering. New Delhi, 3; 1977. p. 2157–62. Hazirbaba K. Pore pressure generation characteristics of sands and silty sand: A strain approach. University of Texas at Austin; 2005. p. 232 Ph.D. dissertation, http://repositories.lib.utexas.edu/bitstream/handle/2152/1791/hazirba bak18803.pdf. Hazirbaba K, Rathje EM. Pore pressure generation of silty sands due to induced cyclic shear strains. J Geotech Geoenviron Eng ASCE 2009;135 (12):1892–905. Høeg K, Dyvik R, Sandbækken G. Strength of undisturbed versus reconstituted silt and silty sand specimens. J Geotech Geoenviron Eng ASCE 2000;126 (7):606–17. Jafarzadeh F, Sadeghi H. Experimental study on dynamic properties of sand with emphasis on the degree of saturation. Soil Dyn Earthq Eng 2012;32:26–41. James M, Aubertin M, Wijewickreme D, Wilson GW. A laboratory investigation of the dynamic properties of tailings. Can Geotech J 2011;48:1587–600. Kjellman W. Testing the shear strength of clay in Sweeden. Geotechnique 1951;2(3):225–32. Kuerbis R, Negussey D, and Vaid, YP. Effect of gradation and fines content on the undrained response of sand. In: Proceedings of the hydraulic fill structures, ASCE. Geotechnical Special Publication 21; 1988. p. 330–45. Lade PV, Yamamuro JA. Effects of non-plastic fines on static liquefaction of sands. Can Geotech J 1997;34:918–28. Lade PV, Yamamuro JA. Evaluation of static liquefaction potential of silty sand slopes. Can Geotech J 2011;48(2):247–64. Lade PV, Liggio Jr. CD, Yamamuro JA. Effects of non-plastic fines on minimum and maximum void ratios of sand. Geotech Test J ASTM 1998;21(4):336–47. Monkul, MM. On some of the factors influencing the fines' role on liquefaction of silty sands. In: Proceedings of the state of the art and practice in geotechnical engineering, ASCE. Geotechnical Special Publication 225; 2012. p. 799–08. Monkul MM. Influence of gradation on shear strength and volume change behavior of silty sands. Geomech Eng 2013;5(5):401–17. Monkul, MM and Yamamuro, JA. Influence of densification method on some aspects of undrained silty sand behavior. In: Proceedings of the fifth international conference on recent advances in geotechnical earthquake engineering and soil dynamics. San Diego, California, Paper no: 1.23a (CD-ROM); 2010. isbn:1-887009-15-9. Monkul MM, Yamamuro JA. Influence of silt size and content on liquefaction behavior of sands. Can Geotech J 2011;48:931–42. Monkul MM, Yamamuro JA, Lade PV. Failure, instability, and the second work increment in loose silty sand. Can Geotech J 2011;48:943–55. Monkul MM, Lade PV, Etminan E, Senol A. Compressibility as an indicator of liquefaction potential. Geotech Eng J SEAGS AGSSEA 2014;45(4):73–7. Murthy TG, Loukidis D, Carraro JAH, Prezzi M, Salgado R. Undrained monotonic response of clean and silty sands. Geotechnique 2007;57(3):273–88. Ni Q, Tan TS, Dasari GR, Hight DW. Contribution of fines to the compressive strength of mixed soils. Geotechnique 2004;54(9):561–9. Papadopoulou A, Tika T. The effect of fines on critical state and liquefaction resistance characteristics of non-plastic silty sands. Soils Found JGS 2008;48 (5):713–25. Peacock WH, Seed HB. Sand liquefaction under cyclic loading simple shear conditions. J Geotech Geoenviron Eng ASCE 1968;94(SM3):689–708.

36

M.Murat Monkul et al. / Soil Dynamics and Earthquake Engineering 75 (2015) 27–36

[39] Pitman TD, Robertson PK, Sego DC. Influence of fines on the collapse of loose sands. Can Geotech J 1994;31(5):728–39. [40] Polito CP, Martin II JR. A reconciliation of the effects of non-plastic fines on the liquefaction resistance of sands reported in the literature. Earthq Spectra EERI 2003;19(3):635–51. [41] Porcino D. and Caridi G. Pre- and post-liquefaction response of sand in cyclic simple shear. In: Proceedings of the dynamic response and soil properties, ASCE. Geotechnical Special Publication 160; 2007. p. 1–10. [42] Rad NS, Tumay MT. Factors affecting sand specimen preparation by raining. Geotech Test J 1985;10(1):31–7. [43] Roscoe, KH. An apparatus for the application of simple shear to soil samples. In: Proceedings of the 3rd international conference on soil mechanics and foundation engineering. 1; 1953. p. 186–91. [44] Shen, CK, Vrymoed, JL, Uyeno, CK. The effect of fines on liquefaction of sands. In: Proceedings of the 9th international conference on soil mechanics and foundation engineering. Tokyo, 2; 1977. p. 381–85. [45] Sitharam TG, Dash HK. Effect of non-plastic fines on cyclic behavior of sandy soils. In: Proceedings of the geosustainability and geohazard mitigation, ASCE. Geotechnical Special Publication 178; 2008 p. 319–26. [46] Sivathayalan S, Ha D. Effect of static shear stress on the cyclic resistance of sands in simple shear loading. Can Geotech J 2011;48:1471–84. [47] Sladen JA, Handford G. A potential systematic error in laboratory testing of very loose sands. Can Geotech J 1987;24:462–6. [48] Stewart JP, Chu DB, Seed RB, Ju JW, Perkins WJ, Boulanger RW, et al. Chi-Chi Earthquake reconnaissance report: soil liquefaction. Earthq Spectra 2001;17 (S1):37–60. [49] Taylor ML, Cubrinovski M, Bradley BA. Characteristics of ground conditions in the Christchurch Central Business District. Aust Geomech J 2012;47(4):43–58. [50] Thevanayagam S, Mohan S. Intergranular state variables and stress-strain behaviour of silty sands. Geotechnique 2000;50(1):1–23.

[51] Thevanayagam S, Shenthan T, Mohan S, Liang J. Undrained fragility of clean sands, silty sands and sandy silts”. J Geotech Geoenviron Eng ASCE 2002; 128(10):849–59. [52] Troncoso, JH, Verdugo, R. Silt content and dynamic behavior of tailing sands. In: Proceedings of the 11th international conference on soil mechanics and foundation engineering. San Francisco, 3; 1985. p. 1311–14. [53] Vaid YP, Sivathayalan S. Static and cyclic liquefaction potential of Fraser Delta sand in simple shear and triaxial tests. Can Geotech J 1996;33(2):281–9. [54] Vaid YP, Sivathayalan S, Stedman D. Influence of specimenreconstituting method on the undrained response of sands. Geotech Test J 1999; 22(3):187–95. [55] Wijewickreme D, Sanin MV, Greenaway GR. Cyclic shear response of finegrained mine tailings. Can Geotech J 2005;42:1408–21. [56] Wijewickreme D, Sriskandakumar S, Byrne P. Cyclic loading response of loose air-pluviated Fraser River sand for validation of numerical models simulating centrifuge tests. Can Geotech J 2005;42:550–61. [57] Wood FM, Yamamuro JA, Lade PV. Effect of depositional method on the undrained response of silty sand. Can Geotech J 2008;45:1525–37. [58] Yamamuro JA, Lade PV. Static liquefaction of very loose sands. Can Geotech J 1997;34:905–17. [59] Yamamuro JA, Covert KM. Monotonic and cyclic liquefaction of very loose sands with high silt content. J Geotech Geoenviron Eng ASCE 2001; 127(4):314–24. [60] Yamamuro JA, Wood FM. Effect of depositional method on the undrained behavior and microstructure of sand with silt. Soil Dyn Earthq Eng 2004;24:751–60. [61] Zlatovic, S, Ishihara, K. On the influence of non-plastic fines on residual strength. In: Proceedings of IS-Tokyo ’95, first international conference on earthquake geotechnical engineering. 1; 1995. p. 239–44. [62] Zlatovic S, Ishihara K. Normalized behavior of very loose non-plastic soils: effects of fabric. Soils Found 1997;37(4):47–56.