1-s2.0-s003808061200056x-main (1)

iqra

shear tests

distinguished into liquefaction (cyclic and rapid ow) and residual deformation by comparing both monotonic and cyclic undrained

Earthquake and the 1983 Nihonkai-Chubu Earthquake inJapan) have indicated that extremely large horizontalground deformation can occur in liqueed sandy deposits

The Japanese Geote

www.sciencedjournal homepage: www.els

Soils and Fo

nCorresponding author.

E-mail addresses: [email protected] (G. Chiaro),

Soils and Foundations 2012;52(3):498510in coastal or river areas. When lateral spreading and/orow slides take place, ground displacement may exceedseveral meters, even in gentle slopes with an inclination ofless than a few percent, resulting in severe damage to

0038-0806 & 2012 The Japanese Geotechnical Society. Production and hosting by Elsevier B.V. All rights reserved.

http://dx.doi.org/10.1016/j.sandf.2012.05.008

Peer review under responsibility of The Japanese Geotechnical Society.substantial economical losses as well as a great number of deformation are both poorly understood.Past large-magnitude earthquakes (e.g., the 1964 [email protected] (J. Koseki), [email protected] (T. Sato).behavior. It was found that the presence of initial static shear does not always lead to an increase in the resistance to liquefaction or

strain accumulation; they could either increase or decrease with an increasing initial static shear level depending on the type of loading

pattern and failure behavior. In addition, according to the failure behavior which the specimens exhibited, three modes of development

of large residual deformation were observed.

& 2012 The Japanese Geotechnical Society. Production and hosting by Elsevier B.V. All rights reserved.

Keywords: Large strain; Liquefaction; Static shear stress; Torsional shear tests; Undrained cyclic behavior

Introduction

Slope failure is one of the most serious geotechnicaldisasters brought about by earthquakes that may cause

human losses. Yet, its mechanism is not well understood.In particular, the catastrophic liquefaction-induced failurebehavior of natural and articial slopes of sandy depositsand the consequent development of extremely large groundGabriele Chiaro , Junichi Koseki , Takeshi SatoaCentre for Geomechanics and Railway Engineering, University of Wollongong, Australia

bDepartment of Civil Engineering, University of Tokyo, JapancInstitute of Industrial Science, University of Tokyo, Japan

dIntegrated Geotechnology Institute Ltd., Japan

Available online 15 June 2012

Abstract

This study focused on the role which static shear plays on the large deformation behavior of loose saturated sand during undrained

cyclic loading. A series of undrained cyclic torsional shear tests was performed on saturated Toyoura sand specimens up to single

amplitude shear strain exceeding 50%. Three types of cyclic loading patterns, i.e., stress reversal, intermediate and non-reversal, were

employed by varying the initial static shear level and the cyclic shear stress amplitude. The observed types of failure could bea,b c,n dEffects of initial static shear on lproperties of loose saturated Toyouuefaction and large deformationsand in undrained cyclic torsional

chnical Society

irect.comevier.com/locate/sandf

undations

studies, the shear strain levels employed were limited to the

Foubuildings, infrastructures and lifeline facilities (Hamadaet al., 1994).It is recognized that the behavior of soil elements within

a sloped ground composed of saturated sands is differentfrom that of a level ground during cyclic loading. This isbecause the soil elements are subjected to an initial staticshear stress on the horizontal plane or an assumed failuresurface. During earthquake shaking, these elements aresubjected to additional cyclic shear stress due to shearwaves propagating vertically upward from the bedrock.The superimposition of static and cyclic shear stress canhave a major effect on the response of the soil, leading toliquefaction and the development of extremely largeground deformation.Various studies have focused on the effects of static shear

on the undrained cyclic triaxial behavior of sand. Lee andSeed (1967) and Seed (1968) found that the larger the ratio ofinitial static shear stress to initial conning pressure acting ona horizontal plane, the greater the horizontal cyclic shearstress required to induce liquefaction in a given number ofstress cycles. Furthermore, Vaid and Chern (1983) showedthat the cyclic strength can either increase or decrease due tothe presence of static shear stress and depending on thedifference in density of the specimens, the magnitude of thestatic shear and the denition of liquefaction resistance. Inparticular, for loose sand with higher initial static shear, thecyclic strength was reduced due to ow deformation. Basedon the difference in the effective stress paths and the stressstrain relationships, Hyodo et al. (1991) classied theundrained cyclic behavior of anisotropically consolidatedspecimens into three types, i.e., stress reversal, non-reversaland intermediate. They observed that in the stress reversaland intermediate cases on loose samples, failure could beassociated with liquefaction, while in the non-reversal case,residual deformation brought the sample to failure eventhough no liquefaction had occurred. Failure was notobserved in the non-reversal case on dense specimens. Recentwork by Yang and Sze (2011) involved an investigation ofthe interdependence of major factors affecting the liquefac-tion behavior of sand, such as relative density, conningpressure and static shear. Clearly, the initial static shear stresshas a signicant effect on the liquefaction resistance, which isdependent on the initial relative density and the conningpressure. In addition, there are three different failure modesfor sand under undrained triaxial cyclic loadings, namely,ow-type failure, cyclic mobility and accumulated plasticstrain. Among these, the ow-type failure is the most critical,since it is characterized by abrupt, runaway deformationswith no warning signals.It is recognized that simple shear tests simulate eld

stress conditions expected during earthquakes more accu-rately than triaxial tests. The conclusions achieved byYoshimi and Oh-oka (1975), through the performance ofring shear tests, were substantially opposite to those basedon the triaxial tests by Lee and Seed (1967) and Seed

G. Chiaro et al. / Soils and(1968). They pointed out that to induce liquefaction andthe development of large cyclic shear strain, the reversal ofrange of 1020%. This is due mainly to the mechanicallimitations of the employed apparatus and/or the largeextent of the non-uniform deformation of the specimen athigher strain levels, as well as the technical difcultiesinvolved with correcting the effects of the membrane forceduring the tests.Therefore, it is not possible to fully describe the

occurrence of a liquefaction-induced ground deformationof several meters, which means that ground strain mayreach over 100% on a slightly sloped ground.Based on the above-mentioned background, the aim of

this study is to better understand the role which staticshear plays on the large deformation behavior of loosesaturated sand during undrained cyclic loading. In thispaper, the results of investigations on the effects of theinitial static shear on the undrained cyclic behavior ofsaturated Toyoura sand specimens, subjected to cyclictorsional shear loading up to single amplitude of about50% under various combinations of static and subsequentcyclic shear, are presented.

Test apparatus

To reach extremely large torsional shear displacements, afully automated torque-loading apparatus on hollow cylind-rical specimens (Fig. 1), developed by Koseki et al. (2007)and Kiyota et al. (2008), was employed. It is capable ofachieving double-amplitude torsional shear strain levelsexceeding 100% by using a belt-driven torsional loadingsystem that is connected to an AC servo motor throughelectro-magnetic clutches and a series of reduction gears.A two-component load cell, which is installed inside the

pressure cell, as shown in Fig. 1(a), having torque and axialload capacities of 0.15 kNm and 8 kN, respectively, was usedto measure both the torque and the axial load components.The conning pressure, obtained by the difference in pressurelevels between the cell pressure and the pore water pressure,shear stress is necessary. Vaid and Finn (1979) evaluatedthe cyclic loading behavior of Ottawa sand under planestrain conditions using a simple shear device. They clariedthat, in general, the resistance to liquefaction can eitherincrease or decrease due to the presence of static shear anddepending on the relative density of the specimens, themagnitude of the initial static shear stress and the shearstrain level of interest. Tatsuoka et al. (1982) investigatedthe stressstrain behavior of sand under torsional simpleshear conditions, including the case with static shear. Theirresults were well in accordance with those reported byVaid and Finn (1979), conrming that a torsional simpleshear apparatus could be employed as a very useful toolfor evaluating the cyclic undrained stressstrain behaviorof sand.However, it should be noted that in all of the above

ndations 52 (2012) 498510 499was measured by a high-capacity differential pressure trans-ducer (HCDPT) with a capacity of over 600 kPa. To evaluate

ouG. Chiaro et al. / Soils and F500large torsional deformations, a potentiometer with a wire anda pulley was employed (Fig. 1(b) and (c)).To conduct the cyclic shear tests, the specied shear

stress amplitude was controlled by a computer, whichmonitors the outputs from the load cell, computes theshear stress (i.e., the measured shear stress was correctedfor the effects of the membrane force, as described byKoseki et al., 2005) and controls the device accordingly.

Material, specimen preparation and testing procedures

All the tests were performed on Toyoura sand, whichis uniform sand with a negligible nes content under75 mm (specic gravity GS2.656, maximum void ratioemax0.992, minimum void ratio emin0.632, mean dia-meter D500.16 mm and nes content FC0.1%). Severalspecimens with a relative density in the range of 4448%(i.e., void ratio e0.8330.819) were prepared by the airpluviation method. To minimize the degree of inherentanisotropy in the radial direction of the hollow cylindricalsand specimens, the sample preparation was carried outcarefully by pouring the air-dried sand particles into amold, while moving the nozzle of the pluviator radially andcircumferentially at the same time in alternative directions,

Fig. 1. (a) Torsional shear test apparatus on hollow cylindrical specimen, (b) l

et al., 2008).ndations 52 (2012) 498510i.e., rst in a clockwise direction and then in a counter-clockwise direction (De Silva et al., 2006). In addition, toobtain specimens of highly uniform density, the fallingheight was kept constant throughout the pluviation process.In order to get a high degree of saturation, the double

vacuum method (Ampadu and Tatsuoka, 1993) was emplo-yed; de-aired water was circulated into the specimens and aback pressure of 200 kPa was applied. Skemptons B-values ofmore than 0.96 were observed in all the specimens used inthe tests.The hollow cylindrical specimens, with initial dimensions

of 150 mm in outer diameter, 90 mm in inner diameter and300 mm in height, were isotropically consolidated by increas-ing the effective stress state up to 100 kPa. They were thenmonotonically sheared in order to apply a specied value ofinitial static shear representative of the sloping groundconditions, at a strain rate of 0.5%/min, while keepingdrained conditions. After applying a drained creep loadingfor 5 min or even longer, in order to study the behavior of thesandy specimens under seismic conditions (i.e., liquefactionresistance and/or the development of large deformation),undrained cyclic torsional loading with a constant amplitudeof shear stress was applied at a constant shear strain rate ofabout 2.5%/min.

oading device, and (c) plan view of torque-transmission part (after Kiyota

the combined shear stress achieved zero (tmin0) during theundrained torsional shear loading is called intermediate load-ing (Fig. 2(b)) and the type of loading in which the combinedshear stress was always kept positive is called non-reversal

e: void ratio, Dr: relative density (%) measured at an isotropic stress state

time

0

Fig. 2. Scheme of cyclic torsional shear loadings (adapted from Hyodo

et al., 1991).

FouAs listed in Table 1, cyclic loading tests were performedover a wide range of initial static shear, varying from 0 to25 kPa. Two levels of cyclic shear stress amplitude, 16 kPaand 20 kPa, were employed in this study in order to considervarious combinations of initial static and cyclic shear stress.The loading direction was reversed when the amplitude of thecombined shear stress, which was corrected for the effect ofthe membrane force, reached the target value. During theprocess of the undrained cyclic torsional loading, the verticaldisplacement of the top cap was not allowed, with the aim tosimulate as much as possible the simple shear condition thata ground undergoes during horizontal excitation.It should be noted that the effects of the membrane

of s0c100 kPa,tcyclic: cyclic shear stress (kPa), tstatic: initial static shear stress (kPa),tmaxtstatictcyclic: maximum combined shear stress (kPa),tmintstatictcyclic: minimum combined shear stress (kPa).Table 1

Test conditions.

Test e Dr tcyclic tstatic tmax tmin Loading pattern

1 0.825 46.4 16 0 16 16 Reversal2 0.828 45.5 16 5 21 11 Reversal3 0.824 46.6 16 10 26 6 Reversal4 0.833 44.2 16 15 31 1 Reversal5 0.825 46.5 16 16 32 0 Intermediate6 0.820 47.9 16 17 33 1 Non-reversal7 0.829 45.3 16 20 36 4 Non-reversal8 0.819 48.1 20 0 20 20 Reversal9 0.819 48.0 20 5 25 15 Reversal10 0.828 45.6 20 10 30 10 Reversal11 0.832 44.4 20 15 35 5 Reversal12 0.823 46.9 20 20 40 0 Intermediate13 0.826 46.1 20 25 45 5 Non-reversal

G. Chiaro et al. / Soils andpenetration (MP), due to the excess pore water pressuregeneration, on the liquefaction resistance, was not con-sidered in this study, since their extents would be indepen-dent of the drained static shear applied.

Reversal, intermediate and non-reversal cyclic loading

patterns

The soil elements within a sloped ground are subjected to aninitial static shear stress on the horizontal plane. Duringearthquake shaking, these elements can experience partiallyreversed or non-reversed shear stress loading conditions, dueto the superimposition of the static shear stress with the cyclicshear stress. While referring to Hyodo et al. (1991), three typesof cyclic loading patterns were employed in this study, i.e.,stress reversal, intermediate and non-reversal, as schematicallyshown in Fig. 2. During each cycle of loading in some tests,the combined shear stress value was reversed from positive(tmaxtstatictcyclic40) to negative (tmintstatictcyclico0),or vice versa. This type of loading is hereafter called reversalloading (Fig. 2(a)), whereas the type of loading in which thereversal of the loading direction was made when the value ofMIN>0

CYCLIC

STATIC

MAX

time

MIN

0Fig. 3. Relationships between apparent shear stress due to membrane

force and shear strain.

-20

-20

-10

0

10

20

30S

hear

stre

ss,

(kP

a)

Shear strain, (%)

STATIC= 5 kPa MIN=-11 kPa

-10 0 10 20 30 40 50 60 70

Fig. 4. Typical reversal undrained cyclic torsional shear test results.

ourespectively, and tm and Em are the thickness (0.3 mm)and the Youngs modulus (1492 kPa, after Koseki et al.,2005) of the membrane, respectively.In order to conrm the validity of Eq. (1) in correcting for

the effect of the membrane force, a special test was performedby pouring water between the inner and the outer membranesand shearing the water specimen cyclically under undrainedconditions up to a double-amplitude shear strain of 100%.Fig. 3 shows both the experimental and the theoreticalrelationships between the shear strain and the apparent shearstress that are induced by the membranes due to the torsionaldeformation. The deviation of the actual membrane deforma-tion from the uniform cylindrical one that is assumed in thetheory became larger with an increase in the strain level.Hence, in this study, the shear stress was corrected for theeffect of the membrane force by employing the polynomialapproximation of the measured relationship between g and tm,-60 60-4

-3

-2

-1

0

1

2

3

4

m = 0.06084 +2.9273E-5 2-7.43408E-6 3

Measured

App

aren

t she

ar s

tress

due

to m

embr

anes

, m

(kP

a)

Shear strain, (%)

Undrained torsional shear test on water specimen

DA=100%

Computed

-40 -20 0 20 40

G. Chiaro et al. / Soils and F502shown in Fig. 3.

Undrained cyclic torsional shear behavior of sand with static

shear

As listed in Table 1, undrained cyclic tests were per-formed under reversal, intermediate and non-reversalloading patterns. The typical effective stress paths duringcyclic loading for each type of loading pattern and thecorresponding stressstrain relationships are presented inFigs. 46.As shown in Fig. 4, in the case of reversal loading, cyclic

mobility was observed in the effective stress path, wherethe effective stress recovered repeatedly after reaching thestate of zero effective stress (i.e., full liquefaction). It wasaccompanied by a signicant development of shear strain,as evidenced by the stressstrain relationship.As shown in Fig. 5, in the case of intermediate loading,

the behavior of the specimen was similar to that of thereversal case, in the sense that after achieving a fully0

-20

-10

0

10

20

30

40

Effective mean principal stress, p' (kPa)

Test 2 (Reversal)e=0.828

CYCLIC=16 kPa

STATIC= 5 kPa

MAX=21 kPa

MIN=-11 kPa

40

She

ar s

tress

, (

kPa)

Test 2 (Reversal)e=0.828

CYCLIC=16 kPa MAX=21 kPa

20 40 60 80 100

ndations 52 (2012) 498510liqueed state (p 0), progressive large deformation devel-oped while showing cyclic mobility.Fig. 6 represents the case of non-reversal loading. The state

of zero effective stress was not achieved even after applying208 cycles of loading. Although liquefaction did not occur, alarge shear strain exceeding 50% was reached, and theformation of a spiral shear band could be observed.

Undrained monotonic torsional shear behavior of sand with

static shear

In Fig. 7, the effective stress paths and the stressstrainrelationships during the rst quarter cycle of undrainedloading (i.e., equivalent to undrained monotonic loading)are shown for the three tests. The tests were conductedunder almost the same initial conditions of void ratioand conning pressure by varying the initial static shearlevel.All the specimens initially showed contractive behavior

(i.e., a decrease in the p0 value), during which the shearstress (t) steadily increased to a transient peak (tpeak). The

Fou0-10

0

10

20

30

40

50

60


Test 12 (Intermediate)e=0.823

CYCLIC=20 kPaSTATIC=20 kPa MIN= 0 kPa

MAX=40 kPa

60

She

ar s

tress

, (k

Pa)

Test 12 (Intermediate) CYCLIC=20 kPa MAX=40 kPa

20 40 60 80 100

G. Chiaro et al. / Soils andpeak points mark the initiation of unstable behavior, sincethe shear stress drops with further loading to a transientminimum value (or quasi steady state; Verdugo andIshihara, 1996) during which the specimen deforms undernearly constant shear stress. As soon as the shear stressreaches the phase transformation line (PTL; Ishihara et al.,1975), dilative behavior takes place and the effective stresspaths follow the failure envelope line.Lade (1993) dened the instability line (IL) as the line

that connects the peak points of the effective stress pathsto the origin of the stress space. Furthermore, Kramer(1996) termed this line as the ow liquefaction surface(FLS), since ow liquefaction behavior was observed in thetests in which the monotonic or cyclic loading stress pathexceeds the point of peak stress (tpeak). They showed thatthe slope of this line can be uniquely determined forspecimens having similar void ratios, irrespective of theinitial effective stress level.In the current study, it was found that under the same initial

conditions of void ratio and conning pressure, the larger theinitial static shear stress level, the greater the shear stress at the

-10-10

0

10

20

30

40

50

She

ar s

tress

, (k

Pa)

Shear strain, (%)

e=0.823 STATIC=20 kPa MIN= 0 kPa

0 10 20 30 40 50 60 70

Fig. 5. Typical intermediate undrained cyclic torsional shear test results.0-10

0

10

20

30

40

50

60

STATIC= 25 kPa


Test 13 (Non-reversal)e=0.826

CYCLIC= 20 kPa

MIN= 5 kPa

MAX= 45 kPa

60

She

ar s

tress

, (k

Pa)

= 25 kPaTest 13 (Non-reversal)CYCLIC= 20 kPa MAX= 45 kPa

MIN= 5 kPa

20 40 60 80 100

ndations 52 (2012) 498510 503transient peak state (tpeak). In addition, by drawing a linewhich connects the peak points, a boundary with similarfeatures of the IL or FLS could be dened. In contrast toprevious studies, this line does not pass through the origin ofthe stress space. However, ow liquefaction behavior wasobserved in the tests in which during the rst quarter cycle ofundrained loading the stress path exceeds the correspondingstate of the transient peak stress (tpeak), as was also observedby Kramer (1996).

Failure characteristics of sand with initial static shear by

comparison of monotonic and cyclic undrained behaviors

A comparison between the undrained monotonic behaviorand the cyclic behavior of sand was carried out by Vaid andChern (1985) and Hyodo et al. (1994) using triaxial tests andby Alarcon-Guzman et al. (1988) using torsional shear tests.Vaid and Chern (1985) showed that in cyclic tests, owdeformation may be initiated when the stress path reachesthe critical effective stress ratio line. On the other hand,Alarocn-Guzman et al. (1988) stated that ow deformation

0-10

0

10

20

30

40

50

She

ar s

tress

, (k

Pa)

Shear strain, (%)

STATICe=0.826

10 20 30 40 50 60

Fig. 6. Typical non-reversal undrained cyclic torsional shear test results.

ou00

10

20

30

40

50

PTL' =31

STATIC=20 kPa

STATIC=15 kPae=0.832

e=0.829

STATIC=10 kPae=0.828

She

ar s

tress

, (

kPa)


Instabilityline

PEAK

p' =100 kPa Failure Line

'=34

50p' =100 kPa

20 40 60 80 100 120

G. Chiaro et al. / Soils and F504occurs during cyclic loading when the stress state reaches theeffective path from monotonic tests. Moreover, Hyodo et al.(1994) found the occurrence of ow deformation to betriggered during cyclic loading when the stress state reachesthe softening regions in the effective stress path from mono-tonic tests (i.e., the region between IL and PTL). However,these investigations did not clarify the effects of the initialstatic shear on the modes of failure (i.e., failure due toliquefaction or failure brought about by a large extent ofdeformation; Hyodo et al., 1991) of sand subjected toundrained cyclic loading, which were attempted herein.In this study, the observed types of failure were

distinguished into liquefaction and residual deformationbased on the difference in the effective stress paths and themodes of development of the cyclic residual shear strainduring both monotonic and cyclic loading behavior, asshown in Figs. 810.

Cyclic liquefaction

In some cyclic tests, as typically shown in Fig. 8, theshear stress reached a maximum value (tmax), which was

00

10

20

30

40

STATIC=10 kPa

STATIC=15 kPa

e=0.828

e=0.832STATIC=20 kPa

e=0.829

She

ar s

tress

, (k

Pa)

Shear strain, (%)

PEAK

2 4 6 8

Fig. 7. Typical undrained monotonic torsional shear test results.0

-30

-20

-10

0

10

20

30

40

nFou0-30

-20

-10

0

10

20

30

40

50

MINPEAK

Cycles 1 to 11

MIN>0

CYCLIC=16 kPa

STATIC=20 kPa

PEAK

50

otonic

MAX=36 kPa MIN= 4 kPa

Test 7 (Non-reversal) e=0.829

20 40 60 80 100

ndations 52 (2012) 498510 505shear on the resistance to liquefaction or cyclic strainaccumulation. To address this issue, the liquefactionresistance curves were described in this study in terms ofboth the cyclic stress ratio (CSRtcyclic/p00) and the staticstress ratio (SSRtstatic/p00), as listed in Table 2.Moreover, to describe the liquefaction resistance, the

double-amplitude shear strain (gDA) and/or single-ampli-tude shear strain at the maximum shear stress state (gSA atttmax) are used. In this study, by applying the initialstatic shear, however, the stress conditions become non-symmetric with respect to the initial stress state, asschematically shown in Fig. 11. As a result, gDA is notwell representative of the strain accumulation during cyclicloading. Therefore, in order to be consistent with previousstudies, the resistance against liquefaction (or more strictly,the resistance to strain accumulation) was evaluated interms of the number of cycles required to develop a specicamount of single-amplitude shear strain (gSA).Figs. 1214 show the number of cycles to achieve a

single-amplitude shear strain of gSA=7.5%, gSA=20% andgSA=50%, respectively.

She

ar s

tress

, (k

Pa)

-5 15

-10

0

10

20

30

40

MIN>0

MAX>PEAK

Shear strain, (%)

PEAK

Cycles 1 to 11

Mo

CYCLIC=16 kPa

STATIC=20 kPa

0 5 10

Fig. 10. Typical residual deformation failure behavior.

She

ar s

tress

, (k

Pa)

-2.5-10

0

10

20

30

40

50

RS at =static

0

0

SA (at =max)

cycleends

STATICCYCLIC

CYCLIC

MIN

MAX

Shear strain, (%)

DA

Non-reversal

cyclestarts

loadingunloading

0.0 2.5 5.0 7.5 10.0

ouFig. 12(b) shows that the number of cycles, N7.5, to achievea moderated strain level of gSA7.5%, which would corre-spond to a single-amplitude axial strain of ea5% inundrained cyclic triaxial tests, decreases with an increase inSSR, except for the case of CSR0.16, in which the N7.5value slightly increases to 3.2 at SSR0.160.20 afterachieving a minimum value of N7.51.2 at SSR0.15.On the other hand, Fig. 13(b) reveals that the number of

cycles, N20, to achieve a large shear strain level ofgSA20% rst decreases and then increases with anincrease in SSR, irrespective of the level of CSR. It should

Table 2

Resistance against strain accumulation and failure characteristics.

Test SSR CSR N7.5(gSA7.5%)

N20(gSA20%)

N50(gSA50%)

Type of

failure

1 0.00 0.16 35 38 48 CLQ

2 0.05 0.16 20 26 33 CLQ

3 0.10 0.16 10 13 20 CLQ

4 0.15 0.16 1.2 3.2 6.9 RFL

5 0.16 0.16 1.9 4.0 13 RFL

6 0.17 0.16 3.2 13 30 RSD

7 0.20 0.16 3.2 46 202 RSD

8 0.00 0.20 3.2 6.3 18 CLQ

9 0.05 0.20 2.2 4.4 14 CLQ

10 0.10 0.20 1.1 2.0 6.9 RFL

11 0.15 0.20 1.1 2.0 5.7 RFL

12 0.20 0.20 0.9 2.6 7.8 RFL

13 0.25 0.20 0.9 39 225 RSD

p00 initial effective mean principal stress (100 kPa),SSRtstatic/p00: static stress ratio, CSRtcyclic/p00: cyclic stress ratio,CLQ: cyclic liquefaction, RFL: rapid ow liquefaction, RSD: residual

deformation failure.

G. Chiaro et al. / Soils and F506be noted that the cyclic strain accumulation resistanceshown in Fig. 13 is free from the effects of strainlocalization during undrained cyclic shearing, which mayinitiate at strain levels of about gSA2328%, as evaluatedby Chiaro et al. (2011).Finally, Fig. 14(b) shows that the number of cycles, N50, to

achieve an extremely large shear strain level of gSA50%,has the same characteristics as N20 dened at gSA20%, inthe sense that they can either increase or decrease with anincrease in SSR. However, these relationships between CSRor SSR and N50 should be taken only as reference data, sincethey are affected by strain localization (i.e., the formation ofshear bands) during undrained torsional shear loading(Chiaro et al., 2011).Thus, these test results show that the level of shear strain at

which the resistance against strain accumulation is dened (i.e.,moderate or large strain levels) plays an important role in theevaluation of the effect of the initial static shear on the strainaccumulation resistance characteristics. In addition, the two-phase change in strain accumulation behavior (i.e., rst adecrease and then an increase in strain accumulation resistancewith initial static shear) can be associated with a three-phasechange in failure behavior, namely, from cyclic liquefaction torapid ow liquefaction to residual deformation failure.-40-20

-10

0

10

20

30

unloading

0

0

SA at =max

RS at =

-20 0 20 40 60

ndations 52 (2012) 498510Residual deformation development of sand with initial static

shear

The value of gSA, dened at ttmax, may be used toestimate the largest cyclic shear deformation of slopes duringearthquakes. On the other hand, the residual deformation ofslopes just after earthquakes can be estimated using theresidual shear strain dened at a cyclic shear stress of zero(i.e., ttstatic) (Tatsuoka et al., 1982). However, in thecurrent study, it is found that gSA and gRS almost coincidewith each other (Fig. 11). Therefore, to examine the effects ofthe initial static shear on the residual deformation propertiesof saturated loose sand in undrained cyclic torsional sheartests, the residual shear strain was evaluated in terms of gSA.As already described previously, depending on the extent oftstatic and its combination with tcyclic, sand may undergothree different types of behavior, namely, cyclic liquefaction,rapid ow liquefaction and residual deformation failure. InFig. 15, the modes of development of residual deformation

Fig. 11. Denition of shear strain components: (a) stress reversal and

(b) non-reversal loading conditions.

Fou10.12

0.16

0.20

0.24

0.10; 0.15

SA=7.5%

0.17; 0.200.16

0.050.20

0.10

0.15

Cyclic Liquefaction Rapid Flow Liquefaction Residual Deformation

CSR=0.20

CSR=0.16

Number of cycles to achieve SA=7.5%, N7.5

SSR=0

Cyc

lic s

tress

ratio

,CS

R =

C

YC

LIC

/ p'

0

p'0 = 100 kPa

0.20; 0.25

10 100

G. Chiaro et al. / Soils andassociated with each of the observed types of sand behaviorare reported.In the case of either cyclic or rapid ow liquefaction

behavior, Fig. 15(a) and (b), respectively, the higher thetstatic, the lower the number of cycles necessary to reachextremely large residual deformation. In addition, it can beobserved that, following the achievement of the fullliquefaction state (p0 0), large residual deformation devel-oped in just 1015 cycles. However, in the case of cyclicliquefaction behavior, the accumulation of large residualdeformation may occur only after applying several cyclesof loadings, while in the case of rapid ow liquefactionbehavior, it occurs from the rst cycle of loading. Thesetest results clearly highlight the detrimental effect of tstaticin combination with tcyclic, which reduces the number ofcycles up to the onset of liquefaction and signals thecatastrophic development of extremely large residual defor-mation in the post-liquefaction stage.On the other hand, as shown in Fig. 15(c), in the case of

residual deformation behavior, since liquefaction did not

1

0.0

0.1

0.2

0.3 ratherconstant

ratherconstant

Number of cycles to achieve SA=7.5%, N7.5

increase


CSR=0.20

CSR=0.16drop

Sta

tic s

tress

ratio

, SS

R =

S

TATI

C /

p'0

p'0 = 100 kPa

SA=7.5%

10 100

Fig. 12. Cyclic strain accumulation resistance for gSA7.5%: (a) CSRnumber of cycles and (b) SSRnumber of cycles.1 10 1000.12

0.16

0.20

0.24

0.10; 0.15

SA=20%

0.10; 0.170.16

0.20

0.10

0.05

0.15


CSR=0.20

CSR=0.16

Number of cycles to achieve SA=20%, N20

SSR=0

Cyc

lic s

tress

ratio

,CS

R =

C

YC

LIC

/p' 0

p'0 = 100 kPa

0.25

' 0

ndations 52 (2012) 498510 507occur, extremely large residual deformation may beachieved only by applying a large number of cycles.Such tests results would be useful for investigating the

failure mechanism that caused extremely large residualground deformation in liqueed natural sand depositsduring large-magnitude earthquakes (e.g., the 1964 NiigataEarthquake and the 1983 Nihonkai-Chubu Earthquake)that have occurred in Japan during the past decade, and toassess effective countermeasures to minimize the effects ofthe liquefaction-induced ground deformation of naturaland articial sloped grounds.

Discussion

Resistance to strain accumulation of sand based on torsional

shear and triaxial tests with initial shear

Fig. 16 compares the strain accumulation resistance ofloose saturated Toyoura sand obtained in this study byundrained cyclic torsional shear tests with that obtained by

1 10 100

0.0

0.1

0.2

0.3


suddenincrease


CSR=0.20

CSR=0.16sudden

drop

Sta

tic s

tress

ratio

, SS

R =

S

TATI

C/ p

p'0 = 100 kPa

SA=20%

Fig. 13. Cyclic strain accumulation resistance for gSA20%: (a) CSRnumber of cycles and (b) SSRnumber of cycles.

0.170.05lss

ratio

=

CY

CLI

C /p

' 0

1 10 100

0

20

40Full Liquefaction

State (p'=0)

Res

idua

l def

orm

a

Number of cycles

80

oundati5081 10 100 10000.12


p'0 = 100 kPaCyc

Cyclic Liquefactionp'0

=50%ic s

treHyosimsurecleatermand

(a)

(b)

Sta

tic s

tress

ratio

, SS

R =

S

TATI

C /

Fig.

num0.16

0.16

0.20

0.15CSR=0.16,CS

R0.20

0.24

0.10

CSR=0.20SSR=0


SA=50%

0.25

G. Chiaro et al. / Soils and Fdo et al. (1994) by undrained cyclic triaxial tests, underilar initial conditions of relative density, conning pres-as well as applied static and cyclic shear stress. One canrly see that the cyclic responses of sand, measured ins of residual strain (i.e., gRS7.5% for torsional testseRS5% for triaxial tests), are in contrast to each other:

Under torsional shear loading, the cyclic strain resistancerstly decreases with an increase in the initial static shear.As a result, the initial static shear has a detrimental effecton the liquefaction resistance of sand.On the contrary, under triaxial shear loading, anopposite trend was observed, where the cyclic strainresistance rstly increases with an increase in the initialstatic shear. Hence, in this case, the initial static shearseems to be favorable to the liquefaction resistance ofsands. The possible reason is that the soil under cyclictriaxial shearing experiences both extension and com-pression behavior within a single cycle of loading. Forlow values of initial static shear, the extension behavior

1 10 100 1000

0.0

0.1

0.2

0.3


drop

Rapid Flow Liquefaction Residual Deformation

CSR=0.20

CSR=0.16

suddenincrease

p'0 = 100 kPa

SA

14. Cyclic strain accumulation resistance for gSA50%: (a) CSRber of cycles and (b) SSRnumber of cycles.

Fig.

cycli60

80 Test 1 (Reversal) Test 2 (Reversal) Test 3 (Reversal) Test 8 (Reversal) Test 9 (Reversal)

tion

(%)

ons 52 (2012) 498510is predominant, which may cause the soil to liquefyquickly due to the effects of anisotropy. With anincrease in the initial static shear on the triaxialcompression side, the compression behavior predomi-nates and the soil becomes more resistant to liquefac-tion. Therefore, the initial static shear has a benecialeffect on the liquefaction resistance of soil.

1 10 100

0

20

40

60R

esid

ual d

efor

mat

ion

(%)

Test 4 (Reversal) Test 5 (Intermediate) Test 10 (Reversal) Test 11 (Reversal) Test 12 (Intermediate)

Number of cycles

Full Liquefaction State (p'=0)

1 10 100

0

20

40

60

80

Test 6 (Non-Reversal) Test 7 (Non-Reversal) Test 13 (Non-Reversal)

Res

idua

l def

orm

atio

n (%

)

Number of cycles

15. Modes of development of residual deformation during undrained

c torsional loading.

the other hand, in the case of rapid ow liquefaction,

Ref

pFouCastro (1975) and Castro and Poulus (1977) concluded,based on triaxial test results, that strain accumulationresistance increases with an increase in initial static shear ina similar manner to that reported by Hyodo et al. (1994).However, by investigating the effect of axial extensionduring cyclic triaxial tests, they found that the largerdeformation observed in the extension, with respect tocompression for a given deviator stress, does not corre-spond to the eld conditions; therefore, cyclic triaxial testsgenerally overestimate the cyclic deformation that maydevelop in the eld due to liquefaction.In summary, the evaluation of the effect of the initial static

shear on the liquefaction resistance of sand is signicantlyaffected by the testing method employed, and therefore,should be carefully addressed. To this regard, it is well

0.1 1 10 100 1000Cycles for SA=7.5% or RS=5%

Fig. 16. Strain accumulation resistance of loose Toyoura sand by

undrained cyclic torsional shear and triaxial tests with initial static shear.0.00

0.08

0.16

0.24

0.32

Reversal

Triaxial tests by Hyodo et al. (1994) Torsional shear tests by this study

STA

TIC/p

' 0 or

qS

TATI

C/2

Non-reversal' Crecocontriaformundshe

Con

Iliqusheconshewer

(1)0.40

0.48CYCLIC/p'0 or qCYCLIC/2 p'C= 0.16

Toyoura sandDr=40-50%

G. Chiaro et al. / Soils andgnized that simple shear tests can simulate eld stressditions expected during earthquakes more accurately thanxial tests. Hence, torsional simple shear tests, as per-ed in this study, would be a useful tool for bettererstanding and evaluating the effect of the initial staticar on the cyclic undrained behavior of sand.

clusions

n order to evaluate the large deformation behavior andefaction properties of saturated sand with initial staticar stress, a series of undrained cyclic torsional tests wasducted at varying levels of initial static shear and cyclicar stress amplitude. The following main conclusionse obtained.

From the study of failure mechanisms, based on thedifference in the effective stress paths and the modes ofdevelopment of shear strain during both monotonic andcyclic loading behavior, the observed types of failurecould be distinguished into three types, namely, cyclic

E

Am

b

s

Cast

o

Cas

m

1

t

o

De

t

I

G

Ham

l

S

Hyo

s

thear strain localization characteristics of sand in undrained cyclic

orsional shear tests. In: Proceedings of 8th International Conference

n Urban Earthquake Engineering. Tokyo, Japan, pp. 531536.

Silva, L.I.N., Koseki, J., Sato, T., 2006. Effects of different pluviation

echniques on deformation property of hollow cylinder sand specimens.

n: Proceedings of the International Symposium on Geomechanics and

eotechnics of Particulate Media, Ube, Yamaguchi, Japan, pp. 2933.

ada, M., ORourke, T.D., Yoshida, N., 1994. Liquefaction-induced

arge ground displacement. In: Performance of Ground and Soil

tructures during Earthquakes, 13th ICSMFE, pp. 93108.

do, M., Murata, H., Yasufuku, N., Fujii, T., 1991. Undrained cyclicChia

sngineering, ACSE 114 (10), 10891109.

padu, S.K., Tatsuoka, F., 1993. Effects of setting method on the

ehavior of clays in triaxial compression from saturation to undrained

hear. Soils and Foundations 33 (2), 1434.

ro, G., 1975. Liquefaction and cyclic mobility of saturated sand. Journal

f the Geotechnical Engineering Division, ASCE 101 (GT6), 551569.

tro, G., Poulus, S.J., 1977. Factors affecting liquefaction and cyclic

obility. Journal of the Geotechnical Engineering Division, ASCE

03 (GT6), 501551.

ro, G., Kiyota, T., Koseki, J., 2011. Effects of initial static shear onAlar

mduring the rst cycle, full liquefaction and shear strainof a few percent was achieved. In addition, in most ofthe tests, a residual shear strain exceeding 50% wasreached in less than 10 cycles. On the contrary, in thecase of residual deformation failure, extremely largedeformation could be reached after applying a largenumber of cycles of loading, although liquefaction didnot take place.

erences

con-Guzman, A., Leonards, G.A., Chameau, J.L., 1988. Undrained

onotonic and cyclic strength of sands. Journal of Geotechnicalliquefaction, rapid ow liquefaction and residual defor-mation failure. In the case of stress reversal and inter-mediate loadings, failure was associated with fullliquefaction, followed by extremely large deformation inthe post-liquefaction process. On the other hand, in thecase of non-reversal loading, residual deformationbrought the specimen to failure (i.e., the formation ofspiral shear bands), although liquefaction did not occur.

(2) The test results show that the presence of initial staticshear does not always lead to a monotonic change inthe resistance to cyclic shear strain accumulation. Itcan either increase or decrease due to the increase instatic shear, depending on the magnitude of thecombined shear stress, the type of loading, the failurebehavior and also the extent of the shear strain levels atwhich the resistance against strain accumulation isdened.

(3) The mechanisms of residual strain development dependon the failure behavior of the sand. In the case of cyclicliquefaction, the full liquefaction state (p0 0) followedby a sudden development of residual deformation wasachieved after applying several cycles of loading. On

ndations 52 (2012) 498510 509hear strength and residual shear strain of saturated sand by cyclic

riaxial tests. Soils and Foundations 31 (3), 6076.

Hyodo, M., Tanimizu, H., Yasufuku, N., Murata, H., 1994. Undrained

cyclic and monotonic triaxial behavior of saturated loose sand. Soils

and Foundations 34 (1), 1932.

Ishihara, K., Tatsuoka, F., Yasuda, S., 1975. Undrained deformation and

liquefaction of sand under cyclic stresses. Soils and Foundations 15

(1), 2944.

Kiyota, T., Sato, T., Koseki, J., Mohammad, A., 2008. Behavior of

liqueed sands under extremely large strain levels in cyclic torsional

shear tests. Soils and Foundations 48 (5), 727739.

Koseki, J., Yoshida, T., Sato, T., 2005. Liquefaction properties of

Toyoura sand in cyclic torsional shear tests under low conning

stress. Soils and Foundations 45 (5), 103113.

Koseki, J., Kiyota, T., Sato, T., Mohammad, A.M., 2007. Undrained

cyclic torsional shear tests on sand up to extremely large strain levels.

In: International Workshop on Earthquake Hazard and Mitigation,

Guwahati, India, pp. 257263.

Kramer, S.L., 1996. Geotechnical Earthquake Engineering. Prentice-Hall,

New Jersey.

Lade, P.V., 1993. Initiation of static instability in the submarine Nerlerk

berm. Canadian Geotechnical Journal 30 (6), 895904.

Lee, K.L., Seed, H.B., 1967. Dynamic strength of anysotropically

consolidated sand. Journal of the Soil Mechanics and Foundations

Division, ASCE 93 (SM5), 169190.

Seed, H.B., 1968. Land slides during earthquakes due to soil liquefaction.

Journal of the Soil Mechanics and Foundations Division, ASCE 94

(SM5), 10551122.

Tatsuoka, F., Muramatsu, M., Sasaki, T., 1982. Cyclic undrained stress

strain behavior of dense sand by torsional simple shear test. Soils and

Foundations 22 (2), 5569.

Vaid, Y.P., Chern, J.C., 1983. Effects of static shear on resistance to

liquefaction. Soils and Foundations 23 (1), 4760.

Vaid, Y.P., Chern, J.C., 1985. Cyclic and monotonic undrained response

of saturated sands. Advances in the Art of Testing Soils Under Cyclic

Conditions, ACSE Convention, Detroit, USA, pp. 120147.

Vaid, Y.P., Finn, W.D.L., 1979. Static shear and liquefaction potential.

Journal of the Geotechnical Engineering Division, ASCE 105 (GT10),

12331246.

Verdugo, R., Ishihara, K., 1996. The steady state of sandy soils. Soils and

Foundations 36 (2), 8192.

Yang, J., Sze, Y., 2011. Cyclic behavior and resistance of saturated sand

under non-symmetrical loading conditions. Geotechnique 61 (1),

5973.

Yoshimi, Y., Oh-oka, H., 1975. Inuence of degree of shear stress reversal

on the liquefaction potential of saturated sand. Soils and Foundations

15 (3), 2740.

G. Chiaro et al. / Soils and Foundations 52 (2012) 498510510

Effects of initial static shear on liquefaction and large deformation properties of loose saturated Toyoura sand in...IntroductionTest apparatusMaterial, specimen preparation and testing proceduresReversal, intermediate and non-reversal cyclic loading patterns

Tests resultsCorrection of torsional shear stress for membrane forceUndrained cyclic torsional shear behavior of sand with static shearUndrained monotonic torsional shear behavior of sand with static shearFailure characteristics of sand with initial static shear by comparison of monotonic and cyclic undrained behaviorsCyclic liquefactionRapid flow liquefactionResidual deformation failure

Resistance against cyclic strain accumulationResidual deformation development of sand with initial static shear

DiscussionResistance to strain accumulation of sand based on torsional shear and triaxial tests with initial shear

ConclusionsReferences

1-s2.0-s003808061200056x-main (1)

Documents